Atmospheric carbon monoxide (CO) comparison testing is an inter-laboratory quality assurance program that verifies the accuracy, consistency, and metrological traceability of CO measurements in ambient air. Conducted in accordance with ISO/IEC 17025, ISO/IEC 17043, and ISO 13528, these comparison exercises enable environmental monitoring stations, air quality networks, and accredited third-party laboratories to objectively demonstrate measurement competence. This article explains the technical framework, sample types, analytical methods, statistical evaluation criteria, and practical applications of atmospheric CO comparison testing.
Technical Overview
Carbon monoxide is a colorless, odorless, and toxic gas produced primarily by incomplete combustion of fossil fuels and biomass. Because it is a regulated ambient air pollutant with defined limit values in standards such as GB 3095, the reliability of CO measurement data is critical for air quality assessment, pollution source analysis, and regulatory compliance reporting.
Atmospheric carbon monoxide comparison testing refers to the organized, simultaneous analysis of homogeneous and stable CO samples by two or more laboratories or instruments, followed by statistical evaluation of the reported results against a reference or consensus value. The comparison validates whether each participant's measurement system—including the analyzer, calibration chain, operating procedure, and personnel—produces results that are consistent, unbiased, and traceable to national or international measurement standards.
The testing framework is governed by several key reference documents:
- ISO/IEC 17025, which requires laboratories to monitor the validity of results through inter-laboratory comparison or proficiency testing programs.
- ISO/IEC 17043, which specifies general requirements for the competence of providers of proficiency testing schemes.
- ISO 13528, which defines statistical methods for analyzing inter-laboratory comparison data, including assigned values, z-scores, and precision estimates.
- GB 9801, the national method standard for the determination of carbon monoxide in ambient air by non-dispersive infrared spectrometry.
- ISO 4224 and EN 14626, international and European method standards for ambient CO measurement by gas filter correlation and non-dispersive infrared spectroscopy respectively.
A typical comparison exercise covers the ambient-relevant concentration range, from sub-ppm background levels (approximately 0.5 to 2 mg/m³) up to elevated traffic-influenced or episode levels (10 to 50 mg/m³), ensuring that instruments demonstrate acceptable performance across their full working range.
Testing Samples
Certified Reference Gas Mixtures
The most common comparison samples are gravimetrically prepared certified reference gas mixtures of CO in synthetic air or nitrogen, produced and certified in accordance with ISO 6142 or ISO 6143. Multiple concentration levels are supplied—typically a low-level point near the detection limit, mid-range points representing typical urban ambient concentrations, and a high-level span point—so that linearity and bias can be evaluated across the measurement range.
Collected Ambient Air Samples
For comparisons involving real sample matrices, ambient air is collected into inert fluoropolymer sampling bags or passivated stainless-steel canisters. Sample homogeneity and stability are verified by the coordinating body before distribution, since CO concentrations in bags can degrade over time if materials or handling procedures are inadequate. These samples test the complete analytical chain, including sample transport, storage, and pre-treatment.
Zero Gas and Span Gas
High-purity zero air (CO content below 0.1 ppm) and single-point span gases are circulated to verify baseline drift, span drift, and calibration curve integrity at the time of the comparison, complementing the multi-level concentration measurements.
Testing Items
The core evaluated parameters in an atmospheric CO comparison test typically include:
- Determination of CO concentration at multiple assigned levels and comparison with the certified or consensus reference value.
- Zero drift and span drift of continuous analyzers over a defined observation period.
- Repeatability (within-laboratory precision) assessed from replicate determinations.
- Reproducibility (between-laboratory precision) across all participating laboratories.
- Relative error or bias of reported results against the assigned value.
- Linearity of the calibration curve and verification of the effective detection limit.
- Cross-method comparability, for example between non-dispersive infrared, gas filter correlation, and electrochemical techniques.
Testing Methods
Non-Dispersive Infrared Spectroscopy (NDIR)
NDIR is the benchmark reference method specified in GB 9801 and EN 14626. The analyzer measures infrared absorption by CO in a multi-pass optical cell near its characteristic absorption wavelength of approximately 4.6 µm. Modern instruments incorporate pressure and temperature compensation and interference correction for water vapor and carbon dioxide, achieving sub-ppm detection limits and excellent long-term stability suitable for comparison testing.
Gas Filter Correlation Spectroscopy (GFC)
Specified in ISO 4224, the GFC technique alternates a CO-filled and a neutral gas filter in the infrared beam, providing exceptional selectivity against spectral interference from water vapor, carbon dioxide, and hydrocarbons. It is widely used in national air quality monitoring networks and is well suited as a reference technique in comparison programs.
Electrochemical Sensor Method
Portable electrochemical CO sensors offer low cost and field deployability and are frequently included in comparisons as screening-level instruments. Because they are subject to cross-interference, temperature dependence, and gradual sensitivity loss, comparison testing is particularly valuable for qualifying such devices for defined monitoring applications.
Gas Chromatography with Methanizer and FID
For high-sensitivity applications, CO is separated chromatographically, catalytically converted to methane in a methanizer, and detected by flame ionization detection. This technique provides very low detection limits and serves as a high-accuracy reference method for verifying assigned values in comparison exercises.
Statistical Evaluation of Results
Reported concentrations are evaluated using ISO 13528 statistics. The assigned value is derived from the certified concentration or a robust consensus of participant results. Performance is commonly expressed as a z-score, where values of |z| ≤ 2 are satisfactory, 2 < |z| < 3 are questionable, and |z| ≥ 3 are unsatisfactory. Additional metrics such as the En number (|En| ≤ 1 indicating acceptable agreement), relative deviation limits, and Mandel statistics may be applied to assess bias and precision comprehensively.
Applications
Atmospheric CO comparison testing supports quality assurance and regulatory objectives across multiple sectors:
- Quality control of national and regional ambient air quality monitoring networks, ensuring data validity for regulatory reporting against limit values in GB 3095.
- Objective evidence of technical competence for laboratories seeking or maintaining accreditation to ISO/IEC 17025.
- Performance validation and acceptance testing of continuous emission-adjacent and ambient analyzers before network deployment.
- Qualification of portable and low-cost sensor systems for indicative monitoring and screening studies.
- Harmonization of measurement results between different laboratories, instruments, and analytical principles in research programs and standardization work.
- Root-cause investigation and corrective action verification following anomalous monitoring data.
FAQ
What is the purpose of atmospheric carbon monoxide comparison testing?
Its purpose is to objectively verify that a laboratory's CO measurement results are accurate, precise, and comparable to reference values and to the results of peer laboratories. It identifies systematic bias, calibration deficiencies, or procedural errors, and provides documented evidence of measurement validity required by quality systems and accreditation bodies.
What concentration levels are used in a CO comparison test?
Comparison samples typically span the ambient concentration range, from approximately 0.5 mg/m³ background levels up to 50 mg/m³ elevated episode levels. Multiple levels are used so that instrument linearity, low-level sensitivity, and high-level accuracy can all be assessed within a single exercise.
How are comparison results judged as acceptable or not?
The most common criterion is the z-score calculated according to ISO 13528: results with |z| ≤ 2 are satisfactory, 2 < |z| < 3 are warning-level, and |z| ≥ 3 require investigation and corrective action. Relative deviation thresholds against the certified value and En-number evaluation against stated measurement uncertainties are also widely applied.
How often should a laboratory participate in CO comparison testing?
Accredited laboratories generally participate at least once per year, or at the frequency specified in their quality assurance plan and by the requirements of ISO/IEC 17025 monitoring activities. Additional participation is recommended after major instrument installation, relocation, repair, or method change.
Which analytical methods can participate in the comparison?
Any validated method may participate, including NDIR (GB 9801, EN 14626), gas filter correlation (ISO 4224), electrochemical sensors, and chromatographic reference techniques. Comparisons across different principles are encouraged because they reveal method-specific bias and strengthen the overall traceability of ambient CO measurements.