Specimen (Spline) Density Testing Error Analysis: Methods, Error Sources, and Uncertainty Control

Density testing of molded polymer splines (test specimens) to ISO 1183-1, ASTM D792, ASTM D1505, and related standards is a cornerstone of plastics quality control, yet measured values are frequently shifted by avoidable errors. This article presents a systematic error analysis of spline density testing, covering specimen preparation effects, hydrostatic immersion, density gradient column and pyknometer methods, the dominant error sources, and a practical uncertainty evaluation framework for accredited third-party laboratories and in-house QC teams.

Technical Overview

Density is one of the most fundamental physical properties measured on polymer test specimens, commonly referred to as splines or test bars. For materials such as polyethylene, polypropylene, polyamide, and filled thermoplastics, the density of a molded spline is a sensitive indicator of crystallinity, copolymer content, filler loading, void content, and processing consistency. Because many material datasheets specify density tolerances as tight as plus or minus 0.003 g/cm3, even small measurement errors can lead to false acceptance or rejection decisions.

Spline density is typically determined using the family of methods described in ISO 1183-1 (immersion, liquid pyknometer, and titration methods), ISO 1183-2 (density gradient column), ISO 1183-3 (gas pyknometer), ASTM D792 (displacement methods), ASTM D1505 (density gradient technique), and the corresponding GB/T 1033 series of Chinese national standards. Error analysis of the measurement process aims to identify, classify, and quantify gross, systematic, and random errors, and to translate them into a formal measurement uncertainty statement consistent with the principles of the Guide to the Expression of Uncertainty in Measurement (GUM) and the quality requirements of ISO/IEC 17025 for accredited third-party laboratories.

• ISO 1183-1 / ASTM D792 / GB/T 1033.1: immersion (Archimedes) and liquid pyknometer methods
• ISO 1183-2 / ASTM D1505 / GB/T 1033.2: density gradient column method
• ISO 1183-3 / GB/T 1033.3: gas pyknometer method
• ISO 291 / ASTM D618: standard conditioning atmospheres for test specimens
• ISO 294-1, ISO 293, ISO 3167, ASTM D4703, ASTM D638: specimen molding and geometry standards

Testing Samples

Injection-Molded Multipurpose Splines

The most common samples are injection-molded multipurpose test specimens, such as ISO 3167 Type A dumbbell bars produced to ISO 294-1 or ASTM D638 Type I tensile splines. These bars are usually gated at one end, so local packing pressure, shear history, and cooling rate vary along the flow path, which can produce a measurable density gradient from the gate to the free end. For the immersion method, individual test pieces should generally have a mass of at least 1 g so that the resolution of the balance does not become a limiting error source.

Compression-Molded Plaques and Sheets

Compression-molded plaques produced to ISO 293 or ASTM D4703 are frequently used as the reference geometry for density and crystallinity studies. Because semicrystalline resins are cooled in contact with the mold over a longer cycle, plaque density can differ systematically from injection-molded spline density. The specified cooling rate must therefore be recorded and controlled, since it directly determines the crystallinity and hence the density of the specimen.

Coupons Cut from Finished Parts

When density is used to audit molded components, coupons are cut from defined locations of the part. Care must be taken to avoid mixing skin and core material with very different thermal histories, to machine smooth edges that do not trap air bubbles, and to degrease the coupon to remove cutting fluids, release agents, and fingerprints before weighing.

Testing Items

A complete spline density test report normally includes the following measurement items, each of which carries its own characteristic error contributions:

• Density of the spline in g/cm3 or kg/m3, measured on at least three replicate specimens
• Relative density (specific gravity) referenced to water at a stated temperature
• Repeatability of replicate splines within one molding and measurement session
• Batch uniformity, expressed as the density spread across splines from different moldings or cavities
• Expanded measurement uncertainty of the reported density value
• Comparison against the material datasheet tolerance or specification limit

Testing Methods

Hydrostatic Immersion (Archimedes) Method

The specimen is weighed in air to obtain mass ma, then reweighed while fully immersed in a wetting liquid of known density to obtain the apparent mass mL. The density is calculated as the density of the liquid multiplied by ma divided by the difference between ma and mL, with air buoyancy corrections applied as specified in ISO 1183-1 and ASTM D792. Distilled water containing a small amount of wetting agent is used for materials denser than water, while ethanol or surfactant-treated water is used for polyolefins. The liquid temperature is controlled, typically within plus or minus 0.5 degrees Celsius, and the immersion liquid density is corrected to the actual temperature.

Density Gradient Column Method

A vertical column is filled with a stable density gradient of two miscible liquids, such as water and ethanol systems or zinc chloride solutions, and calibrated with certified glass floats spanning the working range. Conditioned spline fragments are introduced into the column, allowed to reach equilibrium depth, and their resting positions are converted to density through the calibrated height-versus-density curve. When the column is held at 23.0 plus or minus 0.1 degrees Celsius as required by ASTM D1505 and ISO 1183-2, sensitivity on the order of 0.0001 g/cm3 is achievable, making this the preferred method for resolving small density differences.

Pyknometer and Gas Displacement Methods

The liquid pyknometer method of ISO 1183-1 is suitable for small granules, powder, or cut chips, while the gas pyknometer method of ISO 1183-3 uses helium displacement and avoids liquid contact entirely. Gas pycnometry measures the skeletal volume, so it excludes open pores that communicate with the surface but still includes closed internal voids, an important distinction when interpreting results for molded splines containing process-induced porosity.

Error Sources Analysis

Specimen-Related Errors

• Internal voids and gas inclusions from under-packing or moisture in the melt lower the measured density of the spline relative to the resin
• Crystallinity variations caused by mold temperature, melt temperature, holding pressure, and cooling rate can shift semicrystalline resin density by 0.001 to 0.01 g/cm3
• Moisture absorption in hygroscopic polymers such as PA6, PA66, PC, and ABS inflates the weighed mass and biases the density upward
• Surface contamination, release agents, fingerprints, and electrostatically adhering dust add uncontrolled mass
• Density gradients along the spline, for example between the gate end and the flow end, make the sampling location a systematic error factor
• Oversized specimens improve balance statistics, while specimens that are too small amplify the relative weighing error

Instrumental Errors

• Balance readability, calibration error, drift, and static charge; a 0.1 mg readability analytical balance is recommended
• Suspension wire diameter, immersion depth changes between weighings, and surface tension acting on the wire emerging from the liquid
• Thermometer calibration error propagating directly into the immersion liquid density value
• In gradient columns: nonlinearity of the gradient, aging and mixing of the liquids, convection currents, vibration, and the certification uncertainty of the calibration floats

Environmental and Procedural Errors

• Liquid temperature: near room temperature, the density of water changes by roughly 0.0002 g/cm3 per degree Celsius, so a 1 degree error in temperature recording directly biases the result
• Air bubbles adhering to the specimen surface are the single most common cause of low-biased immersion results; degassing the liquid, pre-wetting with ethanol, and brushing the specimen under the surface are standard countermeasures
• Air buoyancy: neglecting the correction for air density of approximately 0.0012 g/cm3 introduces an error below 0.0002 g/cm3 for polyolefins but approaching 0.001 g/cm3 for high-density mineral-filled compounds around 2 g/cm3
• Liquid absorption and swelling of the specimen during the immersion weighing sequence
• Air currents, bench vibration, and temperature gradients around the balance
• Insufficient conditioning; splines must be conditioned to ISO 291 or ASTM D618, and hygroscopic materials may additionally require drying to stable mass

Calculation and Reporting Errors

• Using a liquid density value that does not match the actual liquid temperature
• Premature rounding of intermediate masses before applying the density formula
• Applying the simplified formula when the air-buoyancy-corrected formula of ISO 1183-1 is required
• Unit conversion mistakes between g/cm3 and kg/m3
• Reporting more decimal places than justified by the measurement uncertainty

Measurement Uncertainty Evaluation

A rigorous uncertainty budget models density as a function of the mass in air, the apparent mass in liquid, and the temperature-dependent liquid density. Each input uncertainty is multiplied by its sensitivity coefficient, the contributions are combined in quadrature, and the expanded uncertainty U is obtained with a coverage factor of k equals 2, corresponding to approximately 95 percent confidence. For a 5 g polypropylene spline measured on a 0.1 mg balance, the weighing terms typically contribute only a few parts in one hundred thousand, liquid temperature control within 0.5 degrees Celsius contributes around 0.0001 g/cm3, and technique-dependent effects such as residual bubbles and wetting dominate the budget.

As practical benchmarks, well-controlled immersion methods achieve expanded uncertainties of roughly 0.0005 to 0.001 g/cm3, while gradient column methods routinely reach 0.0002 to 0.0005 g/cm3. Laboratories operating under ISO/IEC 17025 should validate these figures with certified reference materials, control charts on a check specimen, and periodic participation in interlaboratory comparisons, and should always report the result together with its expanded uncertainty and the conditioning state of the spline.

Applications

Error-controlled spline density testing supports a wide range of engineering and quality functions across the plastics industry:

• Incoming resin inspection and verification of supplier datasheet density specifications
• Batch-to-batch consistency monitoring of polyethylene and polypropylene grades via statistical process control charts
• Estimation of relative crystallinity changes caused by mold temperature or cooling rate studies
• Detection of voids and porosity in injection-molded parts by comparing spline density with pellet or melt density
• Distinction of copolymer grades, filler content changes, and contamination events
• Support for failure analysis and process validation in automotive, packaging, pipe, and medical device programs

FAQ

What repeatability should be expected for spline density testing?

As a practical benchmark, a single operator using a well-controlled immersion setup should reproduce results on identical splines within about 0.001 g/cm3, while a properly maintained density gradient column can resolve differences of 0.0002 g/cm3 or better. Significantly larger scatter almost always indicates bubbles, temperature drift, or specimen inhomogeneity rather than true material variation.

Why do immersion and gradient column results sometimes disagree?

The two methods interrogate the specimen differently. Air bubbles and incomplete wetting bias immersion results low, while gradient columns are sensitive to column temperature gradients and calibration drift. Small true differences also arise because the gradient column uses small fragments that may not represent the whole spline. Agreement should be judged against the combined expanded uncertainties of both methods, not against a single nominal value.

How much does liquid temperature really matter?

Near 23 degrees Celsius, the density of water changes by approximately 0.0002 g/cm3 per degree. A one degree error in reading or controlling the liquid temperature therefore shifts the calculated density by the same order, which is comparable to the entire uncertainty budget of a good immersion measurement.

Should hygroscopic splines be dried before testing?

Yes. Materials such as polyamide, polycarbonate, and ABS should be dried or conditioned to a documented equilibrium state, for example to ISO 291 or ASTM D618 conditions, and the conditioning state must be stated in the report. Otherwise absorbed moisture adds mass and produces a high-biased density.

How many splines should be measured?

At least three replicate specimens taken from defined locations of the molding should be measured, and their individual values reported along with the mean. For gate-to-end gradient assessment, samples from the gate region and the far end are compared explicitly.

Can density testing detect voids in molded parts?

Yes, as a comparative indicator. If the density of cut splines is measurably lower than the density of fully dense pellets of the same resin, and the difference exceeds the combined measurement uncertainty, internal voids or porosity are the most likely cause and should be confirmed by cross-sectioning.