Bias Evaluation in Cigarette Tar Determination: Standards, Methods, and Quality Control

Bias evaluation in cigarette tar determination is a core quality assurance process that verifies the trueness of nicotine-free dry particulate matter (NFDPM) measurements produced by routine analytical smoking machines. This technical article explains the internationally recognized test standards, such as ISO 4387, ISO 3308, and GB/T 19609, the principal sources of measurement deviation, and the statistical tools used to quantify, monitor, and control bias in tobacco testing laboratories.

Technical Overview

In cigarette smoke analysis, "tar" is formally defined as nicotine-free dry particulate matter (NFDPM). It is not measured directly but calculated as the total particulate matter (TPM) collected on a Cambridge filter pad minus the nicotine and water contents of the pad extract, all determined under standardized machine-smoking conditions. The result is expressed in milligrams per cigarette.

Bias evaluation addresses the systematic component of measurement error. According to ISO 5725-1, accuracy of a measurement method consists of trueness (closeness of agreement between the mean of many results and a reference value) and precision (closeness of agreement between independent results). A laboratory may achieve excellent repeatability while still producing results that are consistently too high or too low. Bias evaluation identifies, quantifies, and controls exactly this kind of systematic offset.

The evaluation is essential for regulatory reporting of tar yields, verification of product conformity, comparison of data between laboratories and manufacturing sites, and compliance with the technical requirements of ISO/IEC 17025, which expects laboratories to estimate bias and to participate in interlaboratory comparisons or proficiency testing where available.

Principal Sources of Measurement Bias

• Puff volume, puff duration, and puff frequency settings drifting from the specified values due to inadequate calibration of the smoking machine.
• Deviation of the puff flow profile from the standard bell-shaped waveform described in ISO 3308.
• Unintended blocking of filter ventilation holes by the cigarette holder or gripper.
• Excursions in conditioning and testing atmosphere (temperature and relative humidity) outside the limits of ISO 3402.
• Cross-flow air velocity around the burning coal deviating from the defined value of approximately 200 mm/s.
• Incomplete extraction of the Cambridge filter pad, evaporation losses during handling, or pad inefficiency.
• Errors in gas chromatographic calibration, internal standard preparation, and integration of nicotine and water peaks.
• Analytical balance drift, incorrect butt-length marking, port-to-port differences, and leakage in the smoke path.

Testing Samples

Commercial Cigarette Samples

Finished cigarette products are sampled according to ISO 8243 or the corresponding national sampling plan. Cartons are selected at random from defined lots, cigarettes with visible defects are rejected, and the samples are marked with the specified butt length before conditioning. A sufficient number of cigarettes is taken to allow replicate determinations and statistical treatment of the results.

Reference (Monitor) Cigarettes

Certified monitor cigarettes with assigned values for TPM, nicotine, water, and NFDPM are smoked alongside routine samples in every analytical run. Because their reference yields are known, they act as control samples: any significant deviation of their measured tar from the assigned value signals a systematic error in the overall determination chain, from smoking machine to gas chromatograph.

Blind Duplicates, Spiked Extracts, and Blank Pads

Blind duplicate cigarettes from the same sample are inserted to check within-run consistency. Spiked pad extracts are used to verify the recovery of nicotine and water in the chromatographic step, while blank Cambridge filter pads confirm that the extraction solvent, glassware, and analytical system are free of contamination and carryover.

Testing Items

The bias evaluation program covers the full set of quantities that enter the tar calculation, together with the statistical indicators used to judge trueness:

• Total particulate matter (TPM), in mg per cigarette, from the mass gain of the Cambridge filter pad.
• Nicotine in the pad extract, in mg per cigarette, determined by gas chromatography.
• Water in the pad extract, in mg per cigarette, determined by gas chromatography.
• Nicotine-free dry particulate matter (tar, NFDPM), in mg per cigarette, calculated as TPM minus nicotine minus water.
• Puff count to the specified butt length and mean puff volume, as supporting machine-performance data.
• Verification of puffing parameters: puff volume (35 mL), duration (2.0 s), and interval (60 s).
• Bias metrics: absolute and percentage deviation from the assigned value, z-scores, En numbers, and control-chart statistics.
• Precision metrics: repeatability (r) and reproducibility (R) limits in accordance with ISO 5725 and the precision data published in ISO 4387.

Testing Methods

Sample Preparation and Conditioning

Cigarettes are sampled per ISO 8243 and conditioned per ISO 3402 at 22 °C and 60% relative humidity for a minimum of 48 hours before smoking. Conditioning is performed in a controlled climatic chamber, and conditioning time is extended until equilibrium is reached. Uniform conditioning is critical because moisture content directly influences burning rate, puff count, and therefore TPM yield.

Machine Smoking and TPM Collection

Samples are smoked on a linear or rotary routine analytical smoking machine operated under the standard conditions of ISO 3308 and ISO 4387: a puff volume of 35 mL, a puff duration of 2.0 seconds, one puff every 60 seconds, and a bell-shaped puff profile. Cigarettes are smoked to the specified butt length, defined as the longer of 74% of the total cigarette length or, for filter products, the filter overwrap length plus 3 mm. TPM is retained on a pre-conditioned 44 mm Cambridge filter pad holder, typically loading a defined number of cigarettes per pad with replicate pads per determination to enable statistical evaluation.

Extraction and Gas Chromatographic Analysis

Each pad is extracted with isopropanol containing a suitable internal standard. Nicotine in the extract is quantified by gas chromatography with flame ionization detection in accordance with ISO 10315, and water is quantified by gas chromatography, typically with a thermal conductivity detector, in accordance with ISO 10362-1. Modern dual-channel systems allow simultaneous determination of both analytes from a single extract. The corresponding Chinese national method series, with GB/T 19609 as the equivalent of ISO 4387 for TPM and NFDPM, follows the same analytical principle.

Calculation of Tar

TPM per cigarette is obtained by dividing the mass gain of the pad by the number of cigarettes smoked on it. Nicotine and water yields are calculated from the chromatographic results using calibration curves and the internal standard. Tar is then computed as: NFDPM = TPM − nicotine − water, and the final result is reported in mg per cigarette, rounded to 0.1 mg.

Statistical Bias Evaluation Protocol

A structured protocol is applied to convert routine quality control data into a formal bias assessment:

• Establish reference values: certified values of monitor cigarettes and assigned values from proficiency testing schemes are used as the accepted reference.
• Run control samples: monitor cigarettes are smoked in every analytical run and plotted on Shewhart mean and range control charts to detect trends, shifts, and out-of-control conditions in real time.
• Estimate the bias: the bias b is calculated as the difference between the laboratory mean and the reference value, and expressed in mg per cigarette and as a relative percentage.
• Test significance: a t-test or comparison against the expanded uncertainty of the reference value determines whether the observed offset is statistically significant or consistent with random variation.
• Use proficiency testing results: z-scores and En numbers from interlaboratory comparisons organized with accredited third-party laboratories provide an independent check of laboratory bias; a modulus of z at or below 2 is conventionally regarded as satisfactory.
• Build the uncertainty budget: significant bias components are incorporated into the measurement uncertainty estimate following ISO 21748, using repeatability, reproducibility, and bias data from validation studies.
• Take corrective action: when bias exceeds acceptance limits, root-cause analysis covers machine calibration, holder condition, chromatographic performance, environmental control, and operator technique, with documented corrective and preventive actions as required by ISO/IEC 17025.

Extended Application under Intense Smoking Conditions

The same bias evaluation framework can be extended to tar determination under the intense smoking regime of ISO 20778, which specifies a 55 mL puff volume, a 30-second puff interval, and 50% blocking of ventilation. Because systematic errors amplify under these conditions, calibration of puff volume and precise ventilation blocking devices are particularly critical control points.

Applications

Bias evaluation of tar determination supports a wide range of technical and regulatory activities across the tobacco industry:

• Regulatory conformity assessment and verification of declared tar yields on packaging in markets that mandate machine-smoking yield reporting.
• Internal quality control of testing laboratories and demonstration of technical competence during ISO/IEC 17025 accreditation assessments.
• Product research and development, including blend, filter, and ventilation design studies that depend on trustworthy comparative yield data.
• Method transfer and comparability studies between different smoking machine models, laboratory benches, or manufacturing and testing sites.
• Instrument qualification, periodic performance verification, and root-cause investigation after equipment maintenance or relocation.
• Market surveillance programs in which independent verification of reported yields by accredited third-party laboratories is required.
• Scientific research on smoke chemistry and smoking regimens, where the trueness of NFDPM data underpins the validity of downstream toxicant-per-mg-tar normalization.

FAQ

What exactly is "tar" in a laboratory context?

Tar is the nicotine-free dry particulate matter (NFDPM) of cigarette smoke collected on a Cambridge filter pad under standard machine-smoking conditions. It equals total particulate matter minus nicotine minus water, and it is reported in mg per cigarette.

How is bias different from precision?

Precision describes the random scatter among repeated results, while bias describes a systematic offset of the mean result from an accepted reference value. A laboratory can be highly precise yet biased; bias evaluation specifically targets and corrects the systematic component.

Which standards govern tar determination and its bias evaluation?

Tar determination follows ISO 4387, supported by ISO 3308 (smoking machine conditions), ISO 3402 (conditioning atmosphere), ISO 8243 (sampling), ISO 10315 (nicotine), and ISO 10362-1 (water). GB/T 19609 is the corresponding Chinese national standard. Bias and precision are evaluated using ISO 5725, ISO 21748, and the quality management requirements of ISO/IEC 17025.

What are the standard machine-smoking conditions?

The ISO regime specifies a 35 mL puff of 2.0 seconds duration taken once every 60 seconds, a bell-shaped puff profile, and smoking to a defined butt length. Any drift in these parameters is a primary source of tar bias.

How is the bias actually quantified?

The bias is calculated as the laboratory mean minus the reference value from monitor cigarettes or proficiency testing. Its significance is judged with a t-test, comparison against the expanded uncertainty, or through z-scores and En numbers from interlaboratory comparisons.

How frequently should bias evaluation be performed?

Monitor cigarettes should be smoked and plotted on control charts in every analytical run, while formal comparisons against certified reference values and participation in interlaboratory proficiency testing are typically performed at least annually or as required by the laboratory's quality system.

What corrective actions are taken when significant bias is detected?

Typical actions include recalibration of puff volume, duration, and flow profile; verification of ventilation blocking; requalification of gas chromatographic calibration standards; review of conditioning chamber performance; leak testing of the smoke path; and retraining of operators, followed by a repeated bias study to confirm the correction.