These sections contain information about: type-— type-approval test conditions, test procedures and test requirements; — analyser performance requirements; — evaluation of the type-appro
General
This European Standard outlines the method for measuring benzene concentration in ambient air using automated sampling and gas chromatography (GC) It details the requirements and components of the GC analyser and its sampling system, along with performance characteristics and minimum criteria A type approval test, consisting of laboratory and field tests, determines the actual performance values for specific analysers The selection of a type approved analyser for field measurements relies on calculating the expanded uncertainty, which considers the analyser's performance characteristics and site-specific conditions The expanded uncertainty must not exceed 25% for fixed measurements or 30% for indicative measurements, as per Directive 2008/50/EC Additionally, the standard provides requirements and recommendations for quality assurance and control in field measurements.
Measuring principle
A measured volume of air is drawn through a sorbent tube, where benzene is retained by suitable sorbents and removed from the air stream The collected benzene is then desorbed using heat and transferred by an inert carrier gas into a gas chromatograph with a capillary column and detector for analysis Before entering the column, the sample can be concentrated on a cryo-trap, which is heated to release the sample, or on a pre-column that removes higher boiling hydrocarbons through back flush.
There are two main types of instruments utilized for sampling: one that samples intermittently during each cycle and another that samples continuously Typical cycle durations range from 15 minutes to 1 hour, as demonstrated in Figures 1 and 2.
Figure 1 — Sampling by single trap
Figure 2 — Sampling by multi-trap
Special attention has to be paid to gases that may co-elute with benzene on the chromatographic column selected, such as hydrocarbons with similar boiling points
The final results for reporting shall be expressed in àg/m 3 using standard conversion factors (see Clause 10).
Type approval test
The type approval test evaluates performance characteristics through a series of prescribed tests This European Standard outlines procedures for determining the actual performance values of at least two analysers in both laboratory and field settings, operated in parallel While laboratory tests exclude the sampling inlet, sampling manifold, and external data acquisition system, they do include the analyser sampling line and filter In contrast, field tests may incorporate a sampling inlet and manifold, but their impact on results must be minimized through proper maintenance.
A qualified body is responsible for conducting these tests The assessment for type approval of an analyzer relies on calculating the expanded uncertainty of the measurement results, which is then compared to a specified maximum uncertainty based on the numerical values of the evaluated performance characteristics.
Type approval of an analyser, along with the subsequent quality assurance (QA) and quality control (QC) procedures, ensures compliance with the data quality requirements specified in Annex I of Directive 2008/50/EC.
Appropriate experimental evidence shall be provided by:
— type approval tests performed under conditions of intended use of the specified method of measurement, and
— calculation of expanded uncertainty of results of measurement by reference to ISO/IEC Guide 98-3:2008.
Field operation and quality control
Before installing and operating a type-approved analyser at a monitoring station, it is essential to conduct an expanded uncertainty calculation using the performance values from type approval tests and the specific conditions of the site This calculation is crucial for demonstrating the analyser's suitability for the actual conditions at the monitoring station.
After the installation of the approved analyser at the monitoring station its correct functioning shall be tested
Quality assurance and quality control requirements are essential for the effective operation and maintenance of the sampling system and the analyser These measures are crucial to ensure that the uncertainty of measurement results obtained in the field remains uncompromised.
General
The installation of the analyser at a monitoring station can involve either a single sampling line or a common sampling inlet connected to a sampling manifold for multiple analysers and equipment The design and conditions of the sampling system significantly affect measurement uncertainty To reduce this uncertainty, specific requirements for the sampling equipment are outlined in the subsequent subclauses.
NOTE In Annex B, different arrangements of the sampling equipment are schematically presented.
Sampling location
The specific location for ambient air sampling and analysis varies based on the type of monitoring station, whether in a rural or urban background area Detailed guidance on micro-scale sampling points is provided in Annex III of Directive 2008/50/EC.
Sampling system
The sampling system shall include a sampling inlet and may include the following components:
— a particle filter placed between the sampling line or manifold and the analyser;
— a sampling pump in case a sampling manifold is used
The design of the sample inlet must effectively prevent rainwater from entering the sampling line or manifold Additionally, the sampling line or manifold should be kept as short as possible to reduce residence time to within 6 seconds, as specified in Annex D.
When utilizing a sampling manifold, it is essential to have an additional pump with adequate capacity to meet the sampling requirements outlined in the previous sections (refer to 6.5 and Annex D).
The material of the sample inlet as well as the sampling line or manifold can influence the composition of the sample These shall be chemically inert to benzene
NOTE In practice, the best materials to be used are borosilicate glass and stainless steel
The sampling line or manifold may be moderately heated to avoid condensation Condensation may occur in the case of high ambient temperature and/or humidity
The influence on the measured concentrations of the pressure drop along the sampling inlet and line or manifold and the particle filter shall be ≤ 1,0 %
A particle filter will be implemented to safeguard the analyser, ensuring it captures all particles that could affect its performance Additionally, both the filter material and its housing must be chemically inert to benzene.
NOTE 1 A pore size of the filter of 5 àm usually fulfils this requirement
NOTE 2 Suitable materials for the filter housing are for example PTFE, stainless steel, or borosilicate glass
Filters should be changed regularly based on the dust levels at the sampling location, as specified in section 9.7 Additionally, the filter housing must be cleaned at least once a year If the filter becomes overloaded, it can lead to the adsorption of benzene onto particulate matter and result in an increased pressure drop within the sampling line.
The location of the particulate filter can lead to contamination of the sampling system due to dust deposition, potentially resulting in benzene losses To mitigate this issue, the sampling system must be cleaned regularly, as outlined in section 9.4.1, with the cleaning frequency tailored to the specific conditions of the site.
To ensure accurate measurement of benzene concentrations, the sampling system and particle filter must be conditioned during initial installation and after each cleaning This process involves sampling ambient air for a minimum of 30 minutes at the designated sample flow rate to prevent temporary decreases in measured concentrations.
Conditioning periods are excluded from the availability calculation of the analyser during type approval tests Additionally, conditioning can occur in the laboratory prior to installation It is important to note that conditioning during field operations is regarded as standard maintenance, meaning that the concentrations measured during these periods do not need to be factored into data capture calculations or annual averages.
Control and regulation of sample volume
The volume of air sampled into the sampling trap shall be maintained within the specifications of the manufacturer of the analyser.
Sampling pump for the manifold
A sampling manifold requires a pump to effectively sample ambient air and facilitate the suction of this air through the manifold The inlet of the sampling pump should be positioned at the end of the manifold to ensure optimal performance It is crucial that the sampling pump or blower is adequately rated to supply all connected analyzers with the necessary air volume To ensure the pump is functioning correctly, the installation of a flow alarm system is recommended For further details, an example of a sampling manifold can be found in Annex D.
The influence of the pressure drop induced by the manifold sampling pump on the measured concentration shall be ≤ 1,0 %.(see 9.6.3.1)
General
Annex C lists typical types of automated gas-chromatographic analysers together with their operational parameters
The automated gas-chromatographic analyser consists of the principal components that are described in 7.2 to 7.7.
Sampling trap
A standard sampling trap, constructed from stainless steel or borosilicate glass, is filled with a sorbent or a combination of sorbents designed to effectively capture benzene This setup allows for the measurement of benzene concentrations at or below 10% of the annual limit value, using a minimal volume of air for sampling.
A sorbent particle size of 0,18 mm to 0,25 mm (60 mesh to 80 mesh) is recommended
The sampling trap will degrade over time due to its repetitive heating and cooling, and shall be changed in accordance with the manufacturer’s recommendations.
Sampling device
The sampling device used may vary between instruments but shall be able to deliver a known sample volume at standard conditions of temperature and pressure.
Thermal desorption unit
The thermal desorption unit can effectively perform single- or two-stage desorption of benzene from the sampling trap, depending on the instrument used This process involves heating the trap quickly while passing a flow of carrier gas, usually nitrogen To ensure complete transfer of the trapped benzene to a secondary trap, pre-column, or analytical column, it is crucial to optimize the temperature, gas flow rate, and duration of the desorption.
When utilizing a secondary trap or pre-column, it is essential to ensure effective desorption of benzene from these traps Furthermore, the secondary desorption process must be optimized to enhance the chromatographic resolution of benzene, particularly in the presence of potential interferents.
Separation unit
The separation unit includes an analytical column and a heating oven, which work together to effectively separate benzene from potential interferents This process is designed to quantify benzene at concentrations equal to or below 10% of the annual limit value within a sufficient timeframe.
NOTE Effective separation and quantification may require the application of temperature programming This will affect the minimum cycle time between measurements.
Detector
The detector shall be able to allow quantification of benzene at concentrations equal to or lower than 10
Flame ionization and photo-ionization detectors are the most commonly used types for measuring % of the annual limit value While photo-ionization detectors operate without the need for additional gases, they are more susceptible to response drift and necessitate more frequent maintenance.
Data processing system
The analyser will feature software designed for the identification and quantification of benzene in chromatograms This software enables data reprocessing post-acquisition to address issues such as retention window drift, which can result in incorrect component identification.
Data processing and reprocessing may be performed at the site using a computer internal to the analyser, or using an external computer
In case of internal data treatment, the software version should be clearly identified in the test report, as part of the unambiguous AMS designation (see 11.1)
Any data processing system associated by the manufacturer with his AMS should be considered as part of the analyser
8 Type approval of benzene analysers
General
The measurement of benzene concentration in ambient air must adhere to the maximum uncertainty limits established by Annex I of Directive 2008/50/EC.
To achieve an uncertainty of 25% for fixed measurements or 30% for indicative measurements, the analyser must meet specific performance criteria outlined in the standard The evaluation of these performance characteristics is conducted through laboratory and field tests By integrating the values of these characteristics into the expanded uncertainty calculation, it can be determined if the analyser complies with the maximum uncertainty limits set by Annex I of Directive 2008/50/EC.
The type approval test involves assessing performance characteristics through both laboratory and field tests, along with calculating the expanded uncertainty A minimum of two analysers of the same type must undergo laboratory testing, and these same analysers are also required to be tested in the field It is essential that all analysers successfully pass both testing phases.
The type approval procedure shall fulfil the certification requirements laid down in European Standards
The type approval tests for EN 15267-1 and EN 15267-2 must be conducted by a competent body that adheres to internationally accepted standards for test laboratories The approval is granted by or on behalf of the competent authority, ensuring that the testing body demonstrates compliance with these rigorous requirements.
NOTE 1 EN ISO/IEC 17025 is a harmonized internationally accepted standard
NOTE 2 An accreditation by a member body of the European co-operation for Accreditation to
EN ISO/IEC 17025 is a demonstration of compliance
EN 15267-1 outlines the fundamental principles for the certification of automated measuring systems (AMS) used to monitor emissions from stationary sources and assess ambient air quality The certification process involves several sequential stages.
— performance testing of an AMS;
— initial assessment of the AMS manufacturer’s quality management system;
EN 15267-2 covers the supplementary requirements for an AMS manufacturer’s management system to
EN ISO 9001 for the control of design and manufacturing of AMS This European Standard also serves as a reference document for auditing the AMS manufacturer’s management system.
Relevant performance characteristics and performance criteria
The performance characteristics and their related criteria, as outlined in Table 1, are essential for both laboratory and field tests, specifically within the normative ranges defined by this European Standard For type-approval tests on analysers outside these certification ranges, performance criteria in absolute units, such as nmol/mol/K, must be adjusted, while those in relative units (e.g., %) remain unchanged A competent body must determine the values of the performance characteristics in Table 1 during the laboratory and field tests, following the procedures detailed in sections 8.4, 8.5, and Annex A.
Table 1 — Relevant performance characteristics and criteria
No Performance characteristic Symbol Section Lab
Test Field test Performance criterion for benzene
10 % of the level of the annual limit s rz 8.4.4 x ≤ 0,20 àg/m 3
2 Repeatability standard deviation at the level of the annual limit value s r,ct 8.4.4 x ≤ 0,25 àg/m 3
3 Lack of fit (residual from the linear regression function) 8.4.5
3a Largest residual from the linear regression function at concentrations higher than zero r max x ≤ 5,0 % of the measured value
4 Sensitivity coefficient of sample gas pressure b gp 8.4.6 x
5 Sensitivity coefficient of surrounding temperature b st 8.4.7 x ≤ 0,08 àg/m 3 /K
6 Sensitivity coefficient of electrical voltage b V 8.4.8 x ≤ 0,08 àg/m 3 /V
7 Interferents at concentration c t (at a level of the annual limit) 8.4.9
7a H 2 O with concentration 19 mmol/mol b b H2O,ct x ≤ 0,015 àg/m 3 /(mmol/mol)
8 Carry over (memory effect) c m 8.4.10 X ≤ 1,0 àg/m 3 as the response when measuring zero gas after test gas
9 Reproducibility standard deviation under field conditions S r,f 8.5.5 x ≤ 0,25 àg/m 3 as the average of a three month period
Long-term drift at zero
No Performance characteristic Symbol Section Lab
Test Field test Performance criterion for benzene
10b Long-term drift at span level a D l,s 8.5.4 x ≤ 10 % of maximum of certification range
11 Short-term drift at span level a D s,s 8.4.3 X ≤ 2,0 àg/m 3 over 12 h
12 Difference sample/calibration port c Δx sc 8.4.11 X ≤ 1,0 %
13 Period of unattended operation 8.5.6 x 14 Days
The analyser's availability is over 90%, with a span level between 70% and 80% of the certification range Notably, a H₂O concentration of 19 mmol/mol equates to 80% relative humidity at 20 °C and 101.3 kPa If applicable, this information should be excluded from the uncertainty budget.
Design changes (EN 15267-1 and EN 15267-2)
The manufacturer shall evaluate all changes to a type-approved analyser The manufacturer shall document all changes and evaluations in accordance with the requirements of EN 15267-1 and
EN 15267-2 in such a way that they can be audited
NOTE See also EN ISO 9001 [4]
Manufacturers must inform the test laboratory and relevant authorities of any design changes made to the analyser, unless they can provide evidence that the analyser continues to meet the performance standards outlined in the original certificate There are three defined classes of changes to type-approved analysers.
— Type 0: changes that have no measurable influence to the performance of the analyser;
— Type 1: changes that can have an influence on the performance of the analyser, but where subsequent tests prove that such changes do not have a significant influence;
— Type 2: changes that have a significant influence on the performance of the analyser
Type 2 changes to the analyser can significantly impact its performance, leading to discrepancies with the certified performance characteristics In such cases, the test laboratory, in consultation with the Competent Authority, will assess whether supplementary or complete retesting is necessary to uphold type approval Further evaluation will be conducted by the test laboratory in collaboration with the relevant authority to address any identified Type 2 changes.
Procedures for determination of the performance characteristics during the
A qualified entity will conduct the assessment of performance characteristics in a laboratory setting as part of the type approval testing process The materials and equipment utilized in these testing procedures must meet the specifications outlined in this European Standard Testing will be carried out on a minimum of two analyzers.
Before using the analyser, it is essential to adhere to the manufacturer's operating instructions, especially concerning the equipment setup and the required quality and quantity of consumable products.
Before conducting any tests, it is essential to allow the analyser to warm up for the duration specified by the manufacturer If the warm-up time is not indicated, a minimum of 4 hours is recommended to ensure accurate results.
In normal operational conditions, any instances of self-correction, including their frequency and magnitude, must be accessible to the test laboratory Both the auto zero and auto span drift corrections are subject to the same limitations outlined in the performance characteristics (refer to section 8.4.3).
To ensure accurate measurements, the test gas system must be operated for a sufficient duration before initiating tests, allowing the concentrations to stabilize Additionally, the analyser should be tested with the particle filter in place.
During laboratory tests for the type approval the settings of the monitor shall be as the manufacturer requires All settings shall be noted down in the test report
During the test for each individual performance characteristic, the values of the following parameters shall be stable within the specified range given in Table 2
Table 2 — Set points and stability of test parameters
Sample gas pressure Manufacturer’s specification
(except for the sample gas pressure test, see 8.4.6) ± 0,2 kPa
(except for the temperature of the surrounding air test, see 8.4.7) ± 2 K
Electrical voltage a At nominal line voltage and within manufacturer’s specifications (except for the voltage dependence test, see 8.4.8) ± 1 %
The manufacturer's specification for sample volume allows a variation of ± 1% For analyzers that operate on direct current, the type approval test for voltage variation must be conducted within a range of ± 10% of the nominal voltage.
To determine various performance characteristics, test gases that are traceable to internationally accepted standards must be utilized, unless specified otherwise in this standard The methods for generating these test gases are outlined in Table 3.
Table 3 — Methods for preparation of test gases
Cylinder Gas cylinder (benzene gas mixture in nitrogen) EN ISO 6142
EN ISO 6143 Dynamic preparation Dynamic dilution of benzene gas mixtures
− From continuous injection, diffusion, saturation or permeation devices
EN ISO 6145-10 Static dilution Preparation by means of injecting known amounts of benzene into a known volume of air EN ISO 6144
Possible contamination of zero and test gas shall not significantly influence the results of the laboratory tests Therefore the test gases and zero gas shall meet the following specifications:
— tolerances in the concentrations of the test gases used (except for benzene at 10 % of the limit value): ± 10 %; for benzene at the level of 10 % of the limit value: ± 40 %
— maximum permitted expanded uncertainties (95 % confidence) in the concentration of gases used for laboratory tests: 5 %
Specifications of the purity of test gases and zero gas (expressed as absolute value) are given in Table 4 to Table 6
Table 4 — Specification for purity of test gas
Sum of organic compounds other than benzene (see 8.4.9)
NOTE Certain specific organic compounds may be deliberately included in the test gas In this case, the purity specification excludes these
Table 5 — Specification for purity of zero gas for interferents testing
Sum of organic compounds other than benzene (see 8.4.9)
Table 6 — Specification for purity of zero gas for other tests
Sum of organic compounds other than benzene (see 8.4.9)
For optimal results, it is recommended to conduct field tests using a dedicated set of cylinders and zero air generators specifically reserved for testing purposes Ensuring the stability of both the zero air and the test gas for a duration exceeding the test period is crucial.
After the required stabilization period (8.4.2.1), the analyser shall be adjusted at span level (around
70 % to 80 % of the maximum of the certification range of benzene) Perform 5 successive measurements; calculate the average of the last 4 measurements
The analyser must operate continuously under laboratory conditions while analyzing ambient air After 12 hours, test gas is introduced to the analyser Conduct five consecutive measurements and calculate the average of the final four results.
The short-term drift at span level shall be calculated as follows:
The 12-hour drift at span is represented by D s,s in àg/m³ The average concentration of measurements at span level at the beginning of the drift period is denoted as x s,1 in àg/m³, while x s,2 represents the average concentration at the end of the drift period in àg/m³.
D s,s shall comply with the performance criterion in Table 1
After the required stabilization period (8.4.2.1), the analyser shall be adjusted at span level (around
70 % to 80 % of the maximum of the certification range of benzene)
Eleven successive measurements shall be performed at test gas concentrations at 10 % of the level of the annual limit value and at the level of the annual limit value
From the last 10 results of these measurements the repeatability standard deviations (s r) at 10 % of the annual limit value and at the annual limit value shall be calculated according to:
The repeatability standard deviation (\$s_r\$) is expressed in àg/m³ and is calculated using the formula \$s_r = \sqrt{\frac{1}{n-1} \sum_{i=1}^{n} (x_i - \bar{x})^2}\$, where \$x_i\$ represents the result of the ith measurement, \$\bar{x}\$ is the average of 10 measurements, and \$n\$ is the total number of measurements It is essential that \$s_r\$ meets the performance criteria outlined in Table 1, applicable at both 10% of the annual limit value and the full annual limit value.
The detection limit of the analyser is calculated using the repeatability standard deviation at 10% of the annual limit value, in conjunction with the slope of the calibration function determined in section 8.4.5.
= ⋅ , det 3,3 s r z l B (3) where l det is the detection limit of the analyser in àg/m 3 ; s r,z is the repeatability standard deviation at 10 % of the annual limit value in àg/m 3 ;
B is the slope of the calibration function determined according to Annex A using the data from 8.4.5
8.4.5 Lack of fit of linearity of the calibration function
The linearity of the calibration function of the analyser must be evaluated across a range from 0% to 90% of the maximum certification range for benzene, utilizing a minimum of six concentrations, including zero The analyser should be calibrated at approximately 50% of the maximum certification range For each concentration, including zero, at least four individual measurements are required, with the first measurement of each series excluded from the regression function calculation.
The concentrations shall be applied in the following sequence: 50 %, 10 %, 30 %, 5%, 90 % and 0 % After each change in concentration the first measurement result shall be discarded
The uncertainty in the dilution ratios for the applied concentrations shall be less than 1,5 % with respect to each other
The test is designed to accurately determine deviations from a linear function by incorporating multiple concentrations and repetitions It is robust enough to identify non-linearity both from zero to a lower concentration and from that lower concentration to the higher end of the range.
The calculation of the linear regression function and its residuals will follow the guidelines outlined in Annex A All relative residuals must meet the criteria specified in Table 1 The maximum relative residual, denoted as r max, is crucial for demonstrating compliance with type approval requirement (a) This maximum value will also be used in the calculations for type approval requirements (b) and (d) concerning the annual limit value, as detailed in section 8.6.
8.4.6 Sensitivity coefficient to sample gas pressure
Determination of the performance characteristics during the field test
The performance characteristics in the field must be assessed by a qualified organization as part of the type approval test Additionally, the materials and equipment utilized in these testing procedures must meet the standards outlined in this document.
Over a three-month field test, two analyzers are evaluated for their availability during unattended operation, reproducibility in real-world conditions, and long-term drift The testing process is designed to be continuous, with interruptions only permitted in exceptional circumstances, such as operational disruptions or site changes, which must be justified Nonetheless, the cumulative duration of the testing periods must not be less than three months.
The analysers operate simultaneously at a designated monitoring station, ensuring consistent ambient air conditions To accurately assess long-term drift and reproducibility in field conditions, specific operational requirements must be met.
8.5.2 Selection of a monitoring station for the field test
The selection of a monitoring station is based on the following criteria:
— traffic orientated station (≤ 10 m from kerb-side)
— sufficient capacity of the sampling manifold;
— enough room to place two analysers with calibration gases and/or calibration facilities;
— surrounding temperature control for the analysers, climate controlled at (20 ± 4) °C with temperature registration;
Other items that could be considered:
— presence of telemetry/telephone facilities for remote surveillance of the functioning of the equipment;
Following the installation of the analysers at the monitoring station, it is essential to conduct tests to ensure their proper functioning This testing includes verifying the connections to the sampling manifold, assessing sample gas flows, confirming correct operating temperatures, evaluating the response to test gas, and checking data transmission, along with any other factors deemed necessary by the relevant authority.
After verification of the proper functioning the analysers shall be adjusted at a value of around 70 % to
80 % of the maximum of the certification range
During the three-month period the maintenance requirements, by the manufacturer of the analyser shall be followed
During the initial 14 days, zero and test gas measurements must be conducted at least every two days, using a test gas concentration of approximately 70% to 80% of the maximum certification range A minimum of four independent measurements should be taken and recorded Following this period, measurements with zero and test gas will be required at least every two weeks for the duration of the trial.
For a duration of three months, manual span adjustments to the analyser are prohibited, as they can affect the assessment of long-term drift Any measurement data from the analyser will only be mathematically corrected, based on the assumption of a linear drift since the most recent zero or span check.
During field tests, any auto-rescaling or self-correction functions deemed as "normal operational conditions" must be activated The test laboratory should have access to the magnitude of any self-correction Additionally, the auto span drift corrections that occur during unattended operation (long-term drift) must adhere to the same restrictions outlined in the performance characteristics.
To determine various performance characteristics, test gases, specifically air with a specified benzene concentration, must be utilized, adhering to internationally accepted standards unless stated otherwise in this European Standard Table 3 outlines different methods for generating these test gases.
The test gas used in field tests must demonstrate stability that is negligible in relation to the long-term drift criterion outlined in Table 1 Additionally, any potential contamination of the test gas should not significantly affect the outcomes of the field tests Consequently, the test gas must adhere to specific specifications to ensure reliable results.
— Maximum permitted expanded uncertainties in the concentration of gas used for field tests: 5 %
— Specification of the purity of test gas and zero air used for dilution (expressed as absolute values) are given in Tables 4, 5 and 6
For accurate field tests, it is essential to use a dedicated set of cylinders reserved solely for testing purposes Additionally, the stability of the test gas must be ensured for a duration that exceeds the length of the testing period.
After each bi-weekly period, the drift of the analysers under test must be assessed at both zero and span levels according to the specified procedures If the drift exceeds the established performance criteria for either level, the "period of unattended operation" is defined as the total number of weeks until the infringement is observed, minus two weeks For subsequent uncertainty calculations, the "long term drift" values will be based on the span drift observed during the period of unattended operation.
At the beginning of the drift period five measurements are recorded at span level; the results of the first measurement shall be discarded
The long-term drift is calculated as follows:
The drift at zero, denoted as \$D_{l,z}\$, is measured in àg/m³ The average concentration of measurements at zero level at the start of the drift period, immediately following the initial calibration, is represented as \$x_{z,0}\$ in àg/m³ Conversely, \$x_{z,1}\$ indicates the average concentration of measurements at zero level at the conclusion of the drift period, also in àg/m³.
D l,z shall comply with the performance criterion in Table 1
The drift at span concentration \( c_t \) is represented by \( D_{l,s} \) in percentage The average concentration of measurements at the span level at the start of the drift period, immediately following the initial calibration, is denoted as \( x_{s,0} \) in \(\mu g/m^3\) Conversely, \( x_{s,1} \) indicates the average concentration of measurements at the span level at the conclusion of the drift period, also measured in \(\mu g/m^3\).
D l,z is the drift at zero in àg/m 3
D l,s shall comply with the performance criterion in Table 1
NOTE For the determination of a systematic or random drift, a graph with test gas readings can be useful
8.5.5 Reproducibility standard deviation under field conditions
The reproducibility standard deviation under field conditions is calculated from the measured data during the three-month period
The difference in parallel measurements, denoted as Δx f,i, is determined using the formula Δx f,i = x f,1,i – x f,2,i Here, Δx f,i represents the ith difference measured in àg/m³, while x f,1,i and x f,2,i correspond to the ith measurement results from analyser 1 and analyser 2, respectively, also expressed in àg/m³.
The reproducibility standard deviation under field conditions (s r,f) is calculated according to:
The reproducibility standard deviation under field conditions, denoted as \( s_{r,f} \) in \( \text{µg/m}^3 \), is calculated using the formula \( s_{r,f} = \sum x (13) \) Here, \( n \) represents the number of parallel measurements, and \( x_f \) is the average concentration of benzene measured during the field test, also expressed in \( \text{µg/m}^3 \).
The reproducibility standard deviation under field conditions, s r,f, shall comply with the performance criterion in Table 1
Expanded uncertainty calculation for type approval
The type approval of the analyser involves several key steps: First, each individual performance characteristic tested in the laboratory must meet the criteria outlined in Table 1 Second, the expanded uncertainty, derived from standard uncertainties of specific performance characteristics from laboratory tests, must comply with the criteria in Annex I of Directive 2008/50/EC, which specifies a maximum uncertainty of 25% for fixed measurements and 30% for indicative measurements This criterion applies to continuous measurements at the limit value, with relevant performance characteristics and calculation procedures detailed in Annex E Additionally, the performance characteristics tested in the field must also satisfy the criteria in Table 1, and the expanded uncertainty from both laboratory and field tests must adhere to the same standards set forth in Annex I of Directive 2008/50/EC.
The instrument can be type approved when all 4 requirements are met
9 Field operation and ongoing quality control
General
After selecting a type-approved analyser for a specific measuring task, it is essential to assess its suitability at the designated measuring location, following the evaluation process outlined in section 9.2.
The analyser must be installed at a monitoring station to ensure its normal operation is not affected It should be protected from environmental factors such as dust, rain, snow, direct sunlight, and significant temperature changes Typically, an enclosure with temperature control or air conditioning meets these criteria Additionally, in areas where voltage fluctuations are anticipated, incorporating voltage stabilizers for the power supply may be necessary.
Air conditioning systems can negatively impact analyser performance, so it is essential to minimize direct airflow at the analyser and avoid significant temperature fluctuations caused by these systems.
After installation of the analyser at the measuring station the analyser shall be tested for proper operation This is described in 9.3
Some tests may be performed either in the lab or on site
Once the analyser at the designated site meets the EU data quality objectives outlined in Directive 2008/50/EC, it is essential to implement quality assurance and quality control procedures for the continuous monitoring of benzene concentrations These procedures, detailed in section 9.4, ensure that the measured data adhere to the uncertainty requirements specified in Annex I of the directive, which are 25% for fixed measurements and 30% for indicative measurements.
Suitability evaluation
After selecting a type-approved analyser, it is essential to assess its suitability for the specific conditions at the monitoring site, as temperature fluctuations may affect its ability to meet uncertainty requirements Therefore, controlling the temperature of the air around the analyser may be necessary Additionally, the benzene test gases used for calibrating the analyser must be traceable to internationally recognized standards.
9.2.2 Analyser for a monitoring station or task
An expanded uncertainty assessment for the type-approved analyser must be conducted in accordance with section 9.9 The site-specific conditions to be evaluated are outlined in Table 8, while the methodology for the uncertainty assessment is detailed in Annex F.
Table 8 — Site-specific conditions to be evaluated
Sample pressure range The range of sample gas pressures expected during a whole period of a year shall be estimated.
Surrounding air temperature range The range of temperatures expected shall be within the range specified in the type approval test Temperature may be controlled thermostatically.
Line voltage range a The range of line voltages expected shall be within the range in the type approval test Voltage fluctuations may be controlled by means of voltage stabilizers.
H 2 O concentration range The range of hourly-average H 2 O concentrations during a whole period of a year shall be estimated.
The expanded uncertainty of the calibration gas must be considered, encompassing both the inherent uncertainty of the calibration gas and any uncertainty associated with the dilution system, if applicable.
The calibration frequency specified must be utilized to assess the impact of drift For analyzers functioning on direct current, the type approval test for voltage variation should be conducted within a range of ± 10% of the nominal voltage.
As a “default” input the extreme conditions from the type-approval may be used
If the site-specific conditions are outside the conditions for which the analyser is type approved, then either:
— the analyser is subjected to supplementary tests by a competent body under these site-specific conditions;
— the analyser is tested by the network under these site-specific conditions when the number of sites under concern is limited;
— the uncertainty assessment is performed by extrapolation of the conditions under concern
In every instance, the uncertainty must be recalculated, and a report will be generated The analyser must meet the criteria specified in Annex I of Directive 2008/50/EC, which allows for a 25% margin for fixed measurements.
If an analyser meets the requirement of 30% for indicative measurements, it can be installed and utilized at the monitoring station, with the final decision on its usage resting with the National Competent Authority.
If the site-specific conditions align with the type-approved criteria of the analyzer, the uncertainty can be determined based on these conditions It is essential to document all evaluations conducted.
The analyser shall only be used in the tested configuration
EN ISO 14956 gives information about typical levels of air pollutants However, these should be used with proper care for their representativeness for the situations under consideration
Under specific site conditions, compounds like carbon tetrachloride or butanol may be present Therefore, the network managing the analyzer must ensure accurate benzene determination by selecting appropriate separation conditions, including the choice of analytical column and the temperature program for the column.
Initial installation
After setting up the analyser and sampling system at the monitoring station, it is essential to verify their proper functioning The results of these checks must meet the manufacturer's specifications and the standards outlined in this document, including factors like materials used and residence times Documentation of compliance with both the manufacturer's requirements and the established standards is necessary.
During the initial installation a lack of fit check shall be performed according to 9.6.2
The repeatability standard deviation (s r,z) at a concentration of 10% of the annual limit value must be determined through eleven consecutive measurements The calculation of the repeatability standard deviation (s r,z) will be based on the last ten results from these measurements, following Formula (2).
The results shall fulfil the criteria in Table 9
The repeatability standard deviation shall be combined with the slope of the calibration function to calculate the detection limit of the analyser using Formula (3) The detection limit is used:
— as a criterion for peak rejection;
— when processing data as described in 9.8
It is permitted that these tests be carried out in the laboratory directly before installation at the site or at installation at the site
To evaluate the shortest expected lifetime of a particle filter at a specific site type, the loss of benzene will be measured with and without the filter The test gas will have a benzene concentration of approximately 10 µg/m³, and the filter will need to be replaced if the loss exceeds 3% of the benzene concentration.
The test may be performed at a number of monitoring sites representative of other sites in a network
Before the analyser is put into routine operation, a complete test should be conducted Alternatively, the initial installation can include the first test, with subsequent tests carried out as part of routine QA/QC However, this latter approach may result in data loss, which can negatively impact data capture.
NOTE A relatively simple test procedure is the following:
— supply an overflow of test gas to the analyser, passing the filter, using a t-piece;
— measure the concentration of benzene at initial installation;
— after 1 week, again measure the concentration of benzene;
— replace the filter with a new filter and measure the concentration of benzene;
— calculate the loss of benzene from the relative differences in the 2 concentrations;
— repeat the procedure after 2 weeks, 4 weeks, 8 weeks, 16 weeks, etc until the loss of benzene exceeds the criterion of 3 %;
— establish the maximum lifetime of the filter from the calculated losses
To ensure accurate data collection, it is essential to verify the proper functioning of the data logger or computer system that collects concentration measurements from an analyser at a monitoring station Additionally, the transmission of this data to a central computer system must be thoroughly checked These checks should guarantee that the actual concentrations measured by the analyser are accurately recorded in the data collection system.
Subsequently each time parts of the data registration/transmission process are changed, the proper function of the complete process shall be rechecked
All checks on the proper function of the data collection/transmission system(s) shall be documented.
Ongoing quality assurance/quality control
Quality control is essential for maintaining the accuracy of benzene measurements in ambient air during prolonged monitoring Adhering to maintenance, testing, and calibration procedures is crucial for obtaining reliable and traceable air quality data This section outlines the minimum necessary procedures for maintenance, checks, and calibration to ensure the required quality level is upheld.
The materials and equipment utilized in the test procedures must meet the specified standards to ensure they do not significantly affect the outcomes of the tests.
A National Reference Laboratory that implements continuous quality control procedures can effectively demonstrate its adherence to internationally accepted standards for testing laboratories.
It is recommended that other competent bodies that perform ongoing quality control procedures work in compliance with the requirements of internationally accepted standards for test laboratories
NOTE 1 EN ISO/IEC 17025 is a harmonized internationally accepted standard
NOTE 2 An accreditation by a member body of the European co-operation for Accreditation to EN ISO/IEC 17025 is a demonstration of compliance
9.4.2 Frequency of calibrations, checks and maintenance
The checks and calibrations together with their frequency are summarized in Table 9 Criteria are also given for readjustment, calibration or maintenance of the analyser
Users must recognize that when multiple performance characteristics approach their action criteria, it may result in a breach of the measurement uncertainty data quality objective outlined in Annex I of Directive 2008/50/EC, which is set at 25% for fixed measurements and 30% for indicative measurements To determine if such violations are present, the assessment approach for measurement uncertainty detailed in Annex F can be utilized.
Table 9 — Required frequency of calibration, checks and maintenance Calibration, checks and maintenance Section Frequency Action criteria c,d
Calibration of the analyser 9.5.1 At least every year and after repair Result at zero > detection limit
Repeatability at 10 % of the level of the annual limit 9.3 At least every year and after repair 0,2 àg/m 3
Repeatability at span of the analyser 9.5.1 In combination with calibration, using the data from the calibration
Repeatability standard deviation at span: 0,25 àg/m 3
Annual verification of gases for span checks is required, ensuring that the span exceeds 5.0% of the last certified value Additionally, span checks must be conducted biweekly, maintaining a tolerance of ±5.0% from the initial span value or 0.5 µg/m³, whichever is greater when expressed as a concentration.
Lack of fit check (to be performed in laboratory or in field) 9.6.2,
Annex A Within 1 year of the test at initial installation; subsequently:
Within 1 year after test if the lack-of-fit is within 2,0 % to 5,0 %;
Within 3 years if the lack of fit is ≤ 2,0 %;
Lack-of-fit > 5 % or 0,5 àg/m3 whichever is greater (when expressed as a concentration) for concentrations > 0 Lack-of-fit > 0,5 àg/m 3 at zero
1) influence of pressure drop induced by the manifold pump
Influence on measured values must be at least 1% or 0.5 µg/m³, whichever is greater, with a concentration influence of at least 2% Particle filters should be changed annually, ensuring that the response to test gas passing through the filter is 97% or less Additionally, sampling lines must be tested yearly, with a sample loss of no more than 2.0%.
Calibration, checks and maintenance Section Frequency Action criteria c,d
Changing of consumables 9.7.3 As required by manufacturer b As required
Preventive/routine maintenance of components of the analyser
Regular maintenance is essential, with recommendations for checks every 23 to 25 hours, depending on specific site conditions The presence of a particle filter may vary based on the type of analyzer and its installation Immediate corrective actions are required if any action criteria are infringed, and an evaluation of the impact on previously collected measurement data must be conducted for data validation To meet data capture criteria, a trained operator should inspect the data daily Additionally, if multiple performance characteristics approach their action criteria, it may compromise the measurement uncertainty data quality objective outlined in Annex I of Directive 2008/50/EC, necessitating a reassessment of measurement uncertainty in accordance with the guidelines in Annex F.
Calibration of the analyser
Calibration should be conducted annually using zero air and a concentration of approximately 70% to 80% of the certification range to assess analyzer response and drift More frequent calibrations can provide a clearer understanding of drift and overall analyzer performance.
When benzene concentrations at a specific site are significantly lower than the maximum certification range of the analyzer, specifically by a factor of 10 or more, the calibration concentration can be reduced Consequently, the concentrations used for span checks and lack of fit tests should also be proportionally decreased.
Calibration gases must be introduced prior to the filter Following the initial measurement, four consecutive measurements at span concentration should be conducted to calibrate the analyzer Additionally, the repeatability standard deviation (s r) at the span level must be calculated using Formula (2).
The standard deviation shall comply with the performance criterion in Table 9 at the span level
At least every year and after repair
— result of zero air measurement > detection limit;
— span drift beyond the measurement range in use or tolerances set by the user
If servicing includes manual adjustment to the analyser, this shall only be performed by competent personnel, following strict QA/QC procedures to guarantee documentation and traceability of any adjustment
After servicing the analyser shall be recalibrated
For the calibration of the analyser, several methods are available to generate calibration gases In Table
3 the various methods are given
The benzene calibration gases used for analyser calibration must be traceable to internationally accepted standards, with a maximum permitted uncertainty of 5% at a 95% confidence level for ongoing quality control It is essential that these gases differ from those used in span checks Users must also ensure that the uncertainty of the calibration gas does not contribute to the overall uncertainty budget in a manner that exceeds data quality requirements.
The purity of the gases shall be as specified in Table 4 Otherwise, the uncertainty due to the presence of excess impurities shall be included in the uncertainty budget
After calibration the benzene readings of the analyser as logged by the data acquisition system shall be adjusted in accordance with the following formula:
− calgas bz bz bz,zero bz,cal bz,zero
The corrected benzene reading (\$x_{bz}\$) in µg/m³ is calculated using the formula: \$$x_{bz} = y_{bz} \times \left( \frac{x_{calgas}}{x_{bz,cal} - x_{bz,zero}} \right)\$$In this equation, \$x_{calgas}\$ represents the concentration of benzene gas used for calibrating the analyzer, while \$x_{bz,cal}\$ and \$x_{bz,zero}\$ are the analyzer readings during span and zero calibration, respectively The variable \$y_{bz}\$ denotes the analyzer's reading during actual measurements.
Checks
Test gas can be provided through a gas cylinder, an external calibrator unit, or generated internally within the analyzer The concentration of the test gas should be approximately 70% to 80% of either the maximum certification range or the user-defined range.
To ensure the stability of gases used for span checks, verification must occur annually using reference gases that are traceable to internationally accepted standards These reference gases must meet the specifications outlined in Table 4, and the gas utilized for span checks should not deviate by more than 5% from the last certified value.
The purity levels of the benzene gas mixture and zero air for dilution are outlined in Table 4 However, the specifications for impurities, particularly regarding water vapor and organic compounds, can be relaxed In such instances, any uncertainty arising from the presence of excess impurities must be accounted for in the uncertainty budget if it is deemed significant.
Span : ± 5 % of previous verification or 0,5 àg/m 3 whichever is greater (when expressed as a concentration)
Service the span or zero gas generation unit
Test gas shall be measured for a number of times sufficient to get a stable reading, taking into considering the effect of the measurement time on the data capture
The test gas should pass through the particle filter, whenever possible
The differences between two span values obtained from the following formulae shall be calculated to determine if the action criteria have been exceeded:
= S S- 100 x S (16) where Δx s is the difference between the readings of the current span check and the span check after calibration in %;
S i is the reading of the current span check of the analyser;
S 0 is the reading of the most recent span check after calibration of the analyser
At least every two weeks
Span drift ± 5 % or 0,5 àg/m 3 whichever is greater (when expressed as a concentration)
The analyser shall be recalibrated
The lack of fit of the analyser shall be tested using at minimum the following concentrations: 0 %, 10 %,
For benzene certification, measurements should be taken at 50% and 90% of the maximum certification range or within a user-defined range A minimum of three measurements must be conducted at each concentration level, including zero, with the first result being discarded.
— within 1 year of the test at initial installation; subsequently:
— within 1 year after test if the lack-of-fit is within 2,0 % to 5,0 %;
— within 3 years if the lack of fit is ≤ 2,0 %;
> 5,0 % of the measured value or 0,5 àg/m 3 whichever is greater (when expressed as a concentration;
> 0,5 àg/m 3 for residual at zero
Remove analyser from site for further testing and repair if necessary
Lack of fit may be checked either in the laboratory or on site
9.6.3.1 Procedure for measuring pressure drop induced by the manifold pump
Attach the inlet of manometer to sample port on manifold, leave outlet open to atmospheric pressure; record measured pressure drop
The resulting pressure drop should be used to calculate the induced effect on the analyser’s response using the following formula:
The equation \$\Delta P_m g_p m \times \Delta X = b \Delta P 100\$ describes the relationship between the change in the analyser response (\$\Delta X\$) and the pressure drop (\$\Delta P_m\$) caused by the manifold pump Here, \$g_p\$ represents the sensitivity coefficient of the analyser to changes in sample gas pressure, expressed as a percentage of the measured value obtained during laboratory type approval tests.
Influence of the pressure drop induced by the manifold sampling pump on the measured concentration
≤ 1,0 % or 0,5 àg/m 3 whichever is greater (when expressed as a concentration)
Reduce flow through manifold to reduce the induced pressure drop so that the criterion is met
Clean / replace / repair manifold as necessary and re-test
9.6.3.2 Procedure for testing the sample collection efficiency of the sampling system
The flow rate of the test gas in the sampling system must ensure that the residence time is at least equal to that observed during normal operating conditions Typical manifold systems have a specific diameter that influences this flow rate.
~30 mm, length 2 m) have a volume of ~1,5 l
NOTE Individual sampling lines are tested according to 6.3 and 9.6.2
At least every three years
Clean/replace/repair manifold as necessary and re-test
Possible test procedures for the sample collection efficiency include:
— delivering a test gas containing a known concentration of benzene directly to the manifold and
To assess benzene levels in ambient air, introduce a test gas with a known benzene concentration into the sampled air Measure the benzene concentrations at both the top and end of the manifold using two distinct analyzers.
In Annex D an example is given of a possible manifold test performance using the first setup
9.6.4 Treatment of data after exceedance of performance criteria
In the event of a performance criterion violation during checks, it is essential to assess the impact of this violation on the measurement results from the previous to the current check This evaluation aims to enhance time coverage and data capture, with a target of at least 90% for measurements at industrial sites and a minimum of 35% for other locations Additionally, the data capture should account for at least 90% of the measurement time, excluding calibration and routine maintenance periods.
Figure 3 illustrates a flow scheme for evaluating performance and identifying data correction possibilities, highlighting the deviation of performance characteristics from their required values and addressing the annual limit value.
Figure 3 — Flow scheme for performance of evaluation of effects of violation of performance requirements and possibilities for data correction
Checks that would in principle permit application of corrections include: