Interactive comment on “ World Calibration Center for SF 6 – supporting the quality system of the global atmosphere observation ”

1) General comments: The paper describes the inception of a new facility for quality assurance and quality control within the World Meteorological Organization (WMO)/Global Atmosphere Watch (GAW) Programme. In GAW terminology, facilities of this kind are called World Calibration Centre (WCC), regardless if they perform their own primary calibrations or are related to a scale maintained by another laboratory. Among the WCCs for long-lived greenhouse gases a WCC for SF6 has been lacking so far. Therefore the initiative to build up a WCC-SF6 taken by the Korea Meteorological Administration (KMA) together with the Korea Research Institute of Standards and Science (KRISS) is of great value to the GAW Programme with its global monitoring


Introduction
Greenhouse gases (GHGs) have been known as substances causing the greenhouse effect.Most countries including the industrialized nations exert every effort to limit and reduce the greenhouse gas emissions since the Kyoto Protocol's entry into force 2005, which is an agreement to reduce emissions of six greenhouse gases.Accurate observation data of ambient GHGs are vital for the study of the relationship between the global warming and the climate change, but it is not easy to quantify their mixing ratios owing to their globally and temporally insignificant variation.World Meteorological Organization (WMO) GAW programme is operating an international network composed by 80 countries, and ∼ 400 observatory sites to improve scientific knowledge of climate change.
SF 6 in the atmosphere, one of the greenhouse gases regulated by Kyoto protocol, is the rarest among known greenhouse gases.SF 6 can stay in the atmosphere for a long time due to its stability, from 600 yr to 3200 yr.It has higher global warming potential than other greenhouse gases.Recently, the amount of SF 6 mole fraction is reported to be around 7 ppt and being increasing at a rate of 0.22 ppt yr −1 (WMO, 2009b).Therefore, SF 6 observation becomes one of the important activities of the GAW Programme in support of climate change research.According to Global Atmospheric Watch Station Information System (GAWSIS), 37 stations have been observing SF 6 .Most of the stations are located in European and American regions.
The KMA has observed greenhouse gases since 1998 and contributed the GAW programme as a regional GAW station.Currently, we are monitoring CO 2 , CH 4 , N 2 O, SF 6 , and three types of CFCs, namely CFC-11, CFC-12, CFC-113 on a regular basis.
This data are being reported to the World Data Center for Greenhouse Gases (WD-CGG) in Japan.In order to meet the GAW Data Quality Objective (DQO), the KMA was established, the Korea Meteorological Administration (KMA) is now trying to take another step forward to systematically support GAW stations in improving their traceability and quality system for SF 6 , thereby making a contribution to the WMO/GAW.
The KMA proposed to take on the responsibility as World Calibration Center (WCC) for SF 6 within the GAW Programme.
Establishment and application of a standard scale requires expert technologies and several knowhow of gas metrology.As a member of Consultative Committee for Amount of Substance (CCQM) Gas Analysis Working Group (GAWG), KRISS has participated in several international comparison programs since the last 10 yr.In the international comparison CCQM-P41, which was carried out in 2004, the mole fraction scales for carbon dioxide (CO 2 ) were well agreed between WMO and NMIs (van der Veen et al., 2007;Wessel et al., 2008).In the case of methane (CH 4 ), the WMO scale was 1.4 % lower than the NMIs' scale.As the result, WMO changed their official scale to 1.3 % higher against gravimetric scale (Dlugokencky et al., 2005).Recent international comparison of CCQM-K68 had been performed to compare the mole fraction scales for nitrous oxide (N 2 O).The NMIs' average scale was well agreed to WMO scale for N 2 O (Lee et al., 2011).Based on the experience of standard and science, KRISS, NMI of Korea, would like to contribute to the GAW programme by maintaining and distributing the WMO scale of SF 6 to support SF 6 quality.

Scope
Activities of the WCC will involve maintenance of laboratory standards which are traceable to the WMO SF 6 mole fraction scale and audits of GAW sites to ensure the Introduction

Conclusions References
Tables Figures

Back Close
Full observations traceability to the WMO mole fraction scale for SF 6 .According to the strategic plan of the WMO/GAW, in order to implement the GAW QA system WCC is required to conduct regular assessments of stations and apply standard operating procedures.Regular assessments consist of a system audit and a performance audit.The former is to check whether a station complies with the GAW QA system.It involves an assessment of station siting, infrastructures, organizations, operation, etc.The latter is to assess whether the measurement of SF 6 is being carried out in accordance with the GAW DQOs and to check the observation traceability toward the WMO mole fraction scales.The WCC, as one of GAW facilities undertakes the activities for their implementation by cooperating with other GAW facilities and contributes to harmonization of the measurement scale globally and spatially in the long run.

WCC and the WMO SF 6 mole fraction scale
Traditionally, SF 6 mole fraction scales are being provided by three laboratories, National Oceanic and Atmospheric Administration (NOAA), University of Heidelberg and Scripps Institution of Oceanography (SIO).The comparability between these three scales is known as generally good (WMO, 2009b).A need to compare among the SF 6 observation data produced from various GAW stations required one official scale.The scale was developed by NOAA in 2000 and has been prepared by a static gravimetric dilution method from 2006 and maintained up to now.Since 2009, NOAA plays a key role as a SF 6 Central Calibration Laboratory (CCL).

WCC tasks
The WCC's tasks are described in "a WCC in Box 7 of the GAW report no.172 and implemented by shaping the terms of reference in detail as follows; 1. Assist Members operating GAW stations to link their observations to the GAW primary standard.Introduction

Conclusions References
Tables Figures

Back Close
Full The KMA shall provide appropriate assistance for GAW stations under the WMO so that they can produce SF 6 observation results which have traceability to the WMO mole fraction scale which is kept by the CCL, the NOAA/ESRL (Earth System Research Laboratory).Accordingly the KMA shall possess a set of the WMO SF 6 mole fraction scale to ensure their linkage to CCL, to distribute the WMO scale to the GAW stations, and to perform sites audits.For the purpose of the linkage, the KMA obtained from NOAA the WMO mole fraction scale for SF 6 , such as cylinders in the range of 4-15 pmol mol −1 in November 2010.
2. Develop quality control procedures following the recommendations by the scientific advisory groups (SAGs), support the QA of specific measurements and ensure the traceability of these measurements to the corresponding primary standard.
The KMA shall develop procedures to control the quality as recommended by the SAG to support the QA of measurements for SF 6 and ensure the traceability of these measurements within a year after being designated as WCC.Finally, the KMA is willing to cooperate with the NOAA in future for implementation of GAW quality system in measurement of SF 6 .
Because currently the KMA is observing SF 6 with GC/ECD (Gas Chromatograph/Electron Capture Detector), the KMA has an expertise of SF 6 measurement.The analyzing system has been calibrated by the primary standards provided by the KRISS in the past and will be changed into WMO scale from now on.For precise calibration of SF 6 analyzing system, the primary standards with at least 5 different concentration levels are required due to the non-linear characteristics of ECD.
3. Maintain laboratory and transfer standard that are traceable to the primary standard.
In order to maintain laboratory and transfer standards with the traceability to the WMO SF 6 mole fraction scale, the KMA shall establish a system to produce the 7904 Introduction

Conclusions References
Tables Figures

Back Close
Full laboratory standards based on the WMO SF 6 mole fraction scale of the CCL, such as an analysis system as well as a system to collect and compress dry air.
The laboratory standards will be used to calibrate an analysis system, to measure SF 6 in the atmosphere and to produce the transfer or other laboratory standards of various amounts of SF 6 mole fraction which are to be prepared by sampling naturally (or filling artificially) SF 6 sparser or denser air.The laboratory and transfer standard are maintained by calibration against the WMO SF 6 mole fraction scale on a regular basis.
Meantime, compression equipment for dry air, a large number of 30 l and 6 l aluminum cylinders are also necessary.To produce the laboratory standards, the KWA set up a vacuum system with the heating apparatus and a turbo-pump that is able to lower the inside pressure of the cylinder down to less than 10 −3 mbar.
To produce a better laboratory standard KMA will technically cooperate with CCL.
4. Perform regular calibration and performance audits at GAW sites using transfer standards.
The KMA shall carry out basically calibration and audits every 3 yr using SF 6 transfer standards.To do so, RRT and corrective actions will be conducted.Intervals and methods of the performance audits of the stations could be changed if there is any request from the SAG.

Analysis method
A highly repeatable GC method for comparing gas mixtures of SF 6 in air was developed for this study.The comparison analysis is carried out by gas chromatography equipped with µECD (an Agilent 6890N) using an activated alumina-F1 column (2 m × 2 ea, ID 2.0 mm, 80-100 mesh) and a 7 ml volume loop.P-5 (CH 4 5 % in Ar) gas is used as a carrier gas.The carrier gas was introduced with a flow of ∼ 28 ml min −1 depending on the analyzer condition.Samples passes through a Multi Position Valve (MPV) directly connected to each cylinder, then through a mass flow controller (6890N GC/ECD, Agilent, USA) to be finally introduced in the sample loop.A sample flow is optimized to be introduced at a rate of 200 ml min −1 into a loop volume of 7 ml.
Generally two columns, such as propak Q column and Alumina F-1, can separate SF 6 .For this study in order to select an adequate column two different columns were tested.Figure 2 shows two chromatograms by two different columns.In Fig. 2 it is shown that Activated Alumina F-1 column (Restek, USA) separates more effectively SF 6 from oxygen (O 2 ), nitrous oxide (N 2 O) and other gases.It is found that the SF 6 peak from an Alumina F-1 column can be distinguished from others.A propak Q column looks suitable for analyzing N 2 O in air.Therefore for the separation of SF 6 peaks from any others, alumina-F1 column was selected.When we used total 4 m of column, oven temperature 35 • and carrier gas flow 28 ml min −1 , we separated SF 6 clearly from other components in air under the analytical condition even though matrix of the WMO scale was compressed air.
To have a good analytical precision, response of gas chromatogram was examined.Table 1 shows a series of 11 measurement results of two different cylinders of a sample (S) and a reference (R) alternately.For production of one data point from a cylinder at least three injections were carried out and total 30 min analysis run elapsed.One drift during analysis.The drift of the analyzer was within 0.3 % (relative) whose contribution was included in the value of response ratio.The results of the comparison analysis (ISO, 2001b) of sample and reference cylinders are shown in Table 1.The standard deviation of the mean of the results of the 5 ratios (each of which consists of 5 independent comparisons) is 0.22 % (relative) with the maximum deviation of 0.26 %, which comply with the WMO/GAW recommendation on SF 6 .On the basis of this data set and many other similar sets we estimate the standard uncertainty of SF 6 analysis to be 0.3 % (relative).

Calibration to the WMO scale
We received the WMO scale (5 cylinders) from the CCL for SF 6 (NOAA), and analyzed them to calibrate a GC/ECD.The SF 6 measurement results are shown in Table 2 and Fig. 4. The calibration curve obtained from NOAA cylinders fit well with secondary polynomial function of R 2 ∼ 0.99998.Measured mole fractions agree with NOAA's reported values within the standard deviation of each cylinder, which indicates that the analytical ability by KRISS satisfies with WMO recommendation level, ±0.02 ppt.
In Table 2 the differences between the WMO scale and the calibrated value are within ∼ 0.02 ppt of the repeatability uncertainty.Finally, the analyzer calibrated against the cylinders from CCL, because the calibrated results were consistent within DQOs.During the analysis the standard deviation of one data point obtained from 3 or 4 repeats were within 0.02 ppt.

Preparation of working standard -traceable to the WMO scale
In order to maintain laboratory and transfer standards with the traceability to the WMO SF 6 mole fraction scale, the WCC should produce a laboratory standards based on the WMO SF 6 mole fraction scale.
Both systems for analyzing SF 6 (ISO, 2001b) and for sampling a compressed dry air are required to produce a laboratory standard.As an example, of standard production Figures

Back Close
Full of laboratory or transfer, a sampled air (CC315007) was prepared as a working standard first.Usually for air sampling an oil free pump (RIX co.) is used.To remove moisture in air, a chemical trap was connected with air flow line in the front of the pump.Mainly, 29.5 l aluminum cylinder is used for sampling.After preparation of sampled cylinder, a mole fraction of a laboratory or transfer standard is assigned by an analyzer calibrated by the WMO SF 6 mole fraction scale on a regular basis.
To satisfy the WMO DQOs analyzing system should be optimized for the analysis of SF 6 .Figure 4 shows the chromatogram of air sample cylinder (CC315007) for a working standard overlaid with the WMO SF 6 mole fraction scale.As a result of analysis the amount of SF 6 mole fraction was calculated to be 7.52 ± 0.02 ppt on the calibration curve (in Fig. 5) and 7.52±0.02ppt when using two point calibrations (near bracketing), where their uncertainty considered only repeatability.They agreed with each other well within their repeatability uncertainty.
For two points calibration to assign the value and its total uncertainty for an unknown cylinder following equation can be used; where C is an amount of the mole fraction, R is response area of the analyzer, "ref" is reference or standard scale for calibration.From the Eq. ( 1) uncertainty of C unknown can be calculated as; where a u(R i − R j ) means sqrt u(R i ) 2 + u(R j ) 2 .By the above equations finally assigned value can be 7.52 ± 0.05 ppt.Table 3. shows the uncertainty budget and combined uncertainty of the assigned value.7908 Introduction

Conclusions References
Tables Figures

Back Close
Full

Comparison of the scales between WMO and KRISS
We analyzed three NOAA cylinders together with two KRISS cylinders.KRISS cylinders were prepared by gravimetric method (ISO, 2001a) in nitrogen balance.Because the cylinders were prepared with gravimetric method a possible existence of SF 6 in the nitrogen, because it is the main balance gas, is very critical.So a pure nitrogen gas (Deokyang Energen Inc., Korea) was analyzed to check whether SF 6 as an impurity exist.As a measurement result, the amount of SF 6 in pure nitrogen is to be less than 0.01 ppt, when compared to the 6 ppt SF 6 mixture which is prepared in KRISS.In the chromatogram of Fig. 6, blue line is SF 6 of 6 ppt prepared in KRISS and green, pink, and red lines are nitrogen gases used.
As shown in Table 4 and Fig. 6, KRISS scale agreed with the WMO scale within 1.5 %.To reduce the analytical error caused by different matrices (KRISS standard is SF 6 in nitrogen balance and the WMO scale is in air), we put a lot of effort to inject same amount of sample gas into GC/ECD by controlling an injection flow and vent.As reported in 2009 (Min et al., 2009) gas flow rate can be changed when gases with different matrices pass through one thermal mass flow controller, which influences measurement value.We are in preparation of this work.
After all, we found that there is some deviation between two scales which is more than their analysis uncertainties.Therefore we prepare gravimetrically (ISO, 2001a) gas standard mixtures with the almost same balance composition with air such as, 0.78 mol mol −1 N 2 , 0.21 mol mol −1 O 2 and 0.01 mol mol −1 Ar, for more precise and accurate comparison with the WMO scale.

Conclusions
This paper describes the task of the proposed World Calibration Center for SF 6 to implement quality assurance system for SF 6 observations at ambient level which are traceable to WMO scale and satisfies WMO DQO.For the analysis a Gas Introduction

Conclusions References
Tables Figures

Back Close
Full Chromatography/Electron Capture Detector (GC/ECD) was used and calibrated against the WMO scales (five cylinders from the CCL, NOAA, 2011), and consequently a working cylinder which is traceable to the WMO scale was produced.As a result of analysis their uncertainties were < 0.02 ppt for measurement repeatability standard deviation and < 0.02 ppt for analyzer drift.Finally the working standard was consistent with WMO scale within the combined uncertainty of repeatability and drift.
In addition, we found that there is some deviation between two scales of KRISS and WMO.When measured with the same amount injection technique, the WMO scale was 1.5 % higher than that of KRISS.It is regarded that the difference is mainly from different matrix of two mixtures.Accordingly we prepare gravimetrically gas mixtures of the almost same balance composition with air for more precise and accurate comparison with the WMO scale.The results of a production of standard mixture in artificial air will be reported in future paper.Full  Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 5. Provide, in co-operation with the QA/SACs, training and long-term technical help for stations.The KMA shall establish systems and facilities to provide stations with long-term training and technical support regarding QA/QC and measurement in cooperation with the QA/SAC.In particular, the KMA will set up GC/ECD for SF 6 to use it for training within 3 yr after being designated as WCC.Figures Back Close Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | data point comprised of the three repeats has a standard deviation (1 σ) of ∼ 0.20 % (relative).Each response ratio was obtained by dividing sample's peak area by average of both references (before and after sample analysis) to compensate any analyzer Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |
has cooperated with the Korea Research Institute of Standards and Science (KRISS), which is the National Metrology Institute in South Korea and with the Japan Meteorological Administration (JMA), which is the World Calibration Center for CH 4 .Recently, the KMA and KRISS participated in the WMO Round Robin Exercise or Test (RRT) for five greenhouse species.So long as the Central Calibration Laboratory (CCL) for SF 6 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Table 1 .
Comparison data from GC analyzer; one area is from 3 successive analysis.

Table 2 .
Analytical result for SF 6 CRMs from NOAA and KRISS.
a Standard deviation as a value reported by NOAA.b Difference = C calibrated − C WMO-scale .Introduction

Table 3 .
Total uncertainty budget of the working standard (CC315007) of C unknown .

Table 4 .
Measurement result between two scales.