Calibration gases for existing air quality directive pollutants at limit values ( LV )

The European Directive (2008/50/EC) sets up, among other things, the limit values i.e. the maximum allowed concentrations at given time average in the air, for specified pollutants. Calibration of the measurement instruments needs to be performed at regular time intervals. In the framework of an European Joint Research Programme (JRP) named Metrology for Chemical Pollutants in Air (MACPoll) one task aims to provide harmonized dilution methods for air pollutant gases at low concentration for calibration and quality control purposes. The study focuses on the reactive gases nitrogen dioxide and sulphur dioxide at concentration levels corresponding to the limit values given in the European Directive (2008/50/EC). Nitrogen oxide (NO) is studied as well as NO2 because both of them are measured simultaneously for NOx. This work consists in improving the dilution methods for generating calibration standards for SO2, NO, NO2 at limit values and to validate them by an interlaboratory comparison.


Introduction
Historically, the main air pollution problems in both developed and rapidly industrialising countries were typically the high levels of smoke and sulphur dioxide (SO 2 ) emitted following the combustion of sulphurcontaining fossil fuels such as coal, used for domestic and industrial purposes. These days, the major threat to clean air is now posed by traffic emissions. Vehicles with petrol and diesel engines emit a wide variety of pollutants, principally carbon monoxide (CO), nitrogen oxides (NO x ), volatile organic compounds (VOCs) and particulate matter (PM2.5; PM10) which have an increasing impact on urban air quality.
Photochemical reactions resulting from the action of sunlight on nitrogen dioxide (NO 2 ) and VOCs, lead to the formation of ozone (O 3 ). After formation the ozone takes part in the secondary reaction e.g. with NO and impacts rural areas far from the original emission site as a result of the chemical transformation and long range transport..
Because of the potential impacts of pollutants, welfare and the natural environment, ambient concentrations for a number of pollutants are measured with different sampling and analytical methods at a wide range of rural and urban monitoring sites throughout the world.

Objectives
In the framework of a Joint Research Programme (JRP) named Metrology for Chemical Pollutants in Air, one task aims to provide harmonized preparative dilution methods of air pollutant gases for calibration and quality control purposes in air quality monitoring where a need for improved methodologies is present. The work focuses on the reactive gases NO 2 and SO 2 at concentration levels corresponding to the limit values given in the European Directive 2008/50/EC [1] for the measurement of ambient air pollutants. Nitrogen oxide (NO) is studied as well as NO 2 because both of them are measured simultaneously for NOx,. This work consists in improving the methods of dynamic generation of calibration standards for SO 2 , NO, NO 2 at limit values and to validate them by comparison with travelling standards.

Improving dilution methods
Different dilution methods are currently used for the preparation of calibration standards for calibration and quality control purposes. The target in this study is to revise some dilution methods for generating calibration standards of SO 2 , NO and NO 2 at low concentrations which fulfil with the requirements of the European Directive (2008/50/EC) (see Table 1) with traceability to SI units and with an uncertainty of 3%. One hour: 132 nmol/mol (1) One day: 47 nmol/mol (1) Calendar year: 8 nmol/mol (2) NO 2 One hour: 105 nmol/mol (1) Calendar year: 21 nmol/mol (1) (1) European Directive (2008/50/EC) Annex XI/Art. 13 (2) European Directive (2008/50/EC) Annex XIII/Art. 14 The measurement range of 40-150 nmol/mol for SO 2 (the limit value for calendar year is too low) and of 20-100 nmol/mol for NO and NO 2 were selected (there is no limit value for NO, but as NO is normally measured in combination with NO 2 , the work has been focused on NO and NO 2 at the same measurement range).
Four dilution methods are studied : static dilution (SO 2 and NO), permeation method (NO 2 and SO 2 ), dynamic dilution of gas mixtures (NO, NO 2 and SO 2 ) and gas phase titration (NO 2 ) based on the reaction between nitrogen oxide and ozone.
The different source contributions were determined and quantified for each component and for each dilution method between 40 and 150 nmol/mol for SO 2 and between 20 and 100 nmol/mol for NO and NO 2 . The expanded uncertainties calculated for NO, NO 2 and SO 2 on the defined measurement ranges are lower than 3%. The main source of uncertainty comes from the impurities in the dilution gas. However, there are another underscored sources of uncertainty on concentrations of generated calibration standards. In this way each laboratory has identified the different steps of the dilution methods which have to be improved. Some examples of improvements which have been done are given in the following paragraphs.
Concerning the static dilution different types of syringes were tested and the individual syringe volume was determined by an independent method based on the weight of masses. The preparation procedure was automated and the timing was defined for each step to improve the repeatability. The data acquisition system such as the temperature logging of all relevant system parts (vessel, housing, syringes, laboratory) and the pressure logging (vessel, laboratory) was improved.
For the dynamic dilution of high concentration gas mixtures the measurement of the flows (dilution flow and gas flow) was studied by comparing different flow measurement techniques to seek the method that leads to the lowest uncertainty for the dynamic gas mixtures at low concentration such as mass-flow controller, critical orifices and laminar flow element: this study showed that mass-flow controller and laminar flow element led to the lowest uncertainties. The influence of possible reactions between the components of the dynamic gas mixture and different container materials (Teflon (PFA), Sulfinert coatings and stainless steel) was also studied.
Concerning the permeation method the weighing process of permeation tubes and the temperature and pressure stability of the permeation device were improved. As for the dynamic dilution of high concentration gas mixtures the influence of possible reactions between the present components of the dynamic gas mixture and different container materials was also studied.
For gas phase titration facility the stability of the ozone generator as well as the analysis of the ozone represents the major contributions to the uncertainty budget. The analysis of the ozone concentration includes also the uncertainty of the zero air. Different ozone generators have been tested and the contribution of the water concentration in the zero air was studied.

Organization of the interlaboratory comparison 4.1 Determination of the protocol
The comparison consists in measuring the analytical concentrations of dynamic gas mixtures at 2 stable low concentrations for each component in a matrix ''air'' by each laboratory with the improved dilution methods: for SO 2 at 40 nmol/mol and at 150 nmol/mol and for NO, NO 2 at 20 nmol/mol and 100 nmol/mol.
The target will be to obtain a comparability between the results better than 2% for NO and SO 2 and better than 3% for NO 2 .

Choice of the travelling standards
For generating the dynamic gas mixtures two travelling standards have been developed by LNE and by METAS.
A device based on dynamic dilution of gas mixtures with very accurate commercial laminar flowmeters (Molbloc/Molbox) has been set up by LNE for NO and SO 2 (see Figure 1). A gas mixture generator based on the permeation method has been built by METAS for NO 2 (see Figure 2).

Characterization of the travelling standard developed by LNE for NO and SO 2
The travelling standard developed by LNE is composed of two lines : one for the dilution gas and another one for the gas mixtures (at 5 µmol/mol for SO 2 or at 5 µmol/mol for NO) as shown on  For generating the dynamic gas mixtures the flow rates of the high concentration gas mixtures or the dilution gas were optimized. The flow rates have to be sufficiently low not to use a high volume of gas and sufficiently high to avoid purging problems. Different values of flow rates have been tested (see Figure 4) and the results show that the stability time and the concentrations depend on the choice of the flow rates. After optimization it has been decided to set the flow rates at the values given in the Table 2. The choice of the materials used in the device is also very important because the materials (tubings, MFC…) must be adapted to the used gases. For example the stabilisation time of the NO and SO 2 concentrations of the generated gas mixtures depends on the type of tubings used between the dilution system and the analyser (see Figure 5). The Figure 5 shows that the stabilisation is obtained faster with stainless steel with sulfinert treatment and PFA tubings rather than with stainless steel 316 tubings.
As showed in Table 2 the flow rates of the high concentration gas mixtures are low. In this case it's important to reduce the « dead volumes » in contact with the high concentration gas mixture. For reducing the potential « dead volumes » it is used the shortest tubings, a pressure regulator with low volume and without pressure indication and an adapter fitting Type C -1/8 with Sulfinert treatment (low volume) (see Figure 6).  After optimizing the device the reproducibility of the NO and SO 2 concentrations of the dynamic gas mixtures generated with this travelling standard has been evaluated at 20 and 100 nmol/mol for NO and at 40 and 150 nmol/mol for SO 2 . The analysers (chemiluminescence or UV fluorescence) were calibrated with NO and SO 2 gas mixtures at low concentrations certified with LNE's reference standards. NO and SO 2 dynamic gas mixtures have been generated with the travelling standard at the defined flow rates and injected in the analysers. Their analytical concentrations were measured three times on a day and at 5 different days. The reproducibilities calculated for each component and each concentration are given in Table 3.

Characterization of the travelling standard developed by METAS for NO 2
The traceable mobile permeation generator was developed by using mass calibrated permeation devices, an oven with SI-traceable temperature and gas flow measurements [4]. The permeation generator is based on a commercial permeation oven, e.g. Dynacalibrator 150 from VICI Valco Instruments Co. Inc. In order to minimise adsorption of analyte, the stainless steel recipient and the downstream tubing are glass coated. In order to equalise the temperature of the incoming gas the inlet tubing is coiled around the heated chamber. The commercial oven is extended by two controlled gas circuits: a constant carrier gas flow (typically 300 ml/min) through the oven and an adjustable dilution flow. Downstream of the permeation oven the two flows are united. The sum of the two flows as measured by the mass flow meter determines the amount of substance fraction of the generated gas mixture for a given permeation mass flow (see Figure 8). An important requirement for the inside surfaces of the oven chamber and the tubings is that they must be made of the utmost inert materials towards NO 2 and with the lowest possible adsorption rates. Glass largely fulfils these requirements, however the risk of breaking is a clear disadvantage for mobile applications. Stainless steel (SST) with glass coated inner surfaces reveals as an ideal oven material for mobile NO 2 generators.
Another requirement was to minimise the fluctuations of the temperature near or at the permeation device.
To measure the temperature of the permeation device, a calibrated temperature sensor was added and located within the oven at the location of the permeation device. It is a FLUKE Hart 1504 Tweener with incorporated leak tight flexible temperature probe, NTC Thermistor (YSI 46000 series) (see Figure 9). It allows the traceable temperature measurement of the permeation device during the gas generation and thus enables the transfer of the temperature measurement from the micro gravimetric standard to the permeation generator. The drift of the temperature sensor is lower than 0.01°C in 100 months for temperatures below 50°C, its stability is lower than 0.002°C and the resolution of reading instrument is equal to 0.001°C. The following Figure 10 shows the temperature at the permeator and the generated NO 2 concentration over a period of 5 days. Only the MFM for the total gas flow is required to be calibrated. Indicative but stable values for the two MFC's are sufficient. The calibration of the MFM is made with purified or synthetic air and nitrogen against the national primary standard for small gas flows. The reproducibility in the total flow range of the 6 L/min MFM was 0.35 % relative during a time period of 3 years. For CMOS MFM relative reproducibilities better than 0.1 % can be expected over a 6 months period. The instantaneous total gas flow was recorded over a period of 5.5 days (see Figure 11) and the NO 2 concentrations calculated from the actual flow, assuming a constant permeation rate of the permeator. The maximum absolute fluctuation of the gas composition due to the flow variation was 0.25 nmol/mol for a NO 2 concentration close to 380 nmol/mol, corresponding to relative fluctuation equal to 0.06 %.
As for the travelling standard developed for NO and SO 2 by LNE the reproducibility of the permeation generator was estimated by recording the NO 2 concentration generated close to 50 nmol/mol during 14 days (see Figure 12). The results show that the device can be used as travelling standard in the interlaboratory comparison because of its good reproducibility.  Figure 13).

Figure 13.Planning of the interlaboratory comparison
The results of the interlaboratory comparison will be available at the end of 2013 and disseminated at the beginning of 2014.

Conclusion
This European project MACPoll aimed the participants to improve different dilution methods used for generating reference standards for calibration and quality control purposes for ambient air pollutants (NO, NO 2 and SO 2 ) at low concentrations (between 20 and 100 nmol/mol for NO, NO 2 and between 40 and 150 nmol/mol for SO 2 ) which fulfil with the limit values given in the European Directive (2008/50/EC).
Moreover this European project has allowed to develop stable and accurate travelling standards based on permeation for NO 2 and on dynamic dilution of high concentration gas mixtures for NO and SO 2 .
In parallel three guides are in the course of writing for the different dilution methods and will be available at the beginning of 2014: a first guide on dynamic dilution for NO, NO 2 and SO 2 at limit values, a second one on permeation method for NO 2 and SO 2 at limit values and a third one on static dilution for NO and SO 2 at limit values.