Development of a Coulometric Method for Assessing the Concentration of Ambient Levels of CO2/Air in Compressed Gas Mixtures

The coulometric method presented here is a reliable method for the direct analysis of CO2/air cylinder gas mixtures. It is based on Faraday’s laws of electrolysis and therefore no external standardization is required. A series of CO2/air cylinder gas mixtures ranging in concentration from 300 to 375 µmol/mol (ppm) were analyzed and the results compared to those results obtained by non-dispersive infrared (NDIR) analysis with traceability to gravimetric standards. The coulometric method is rapid, sensitive, precise, and with the proper experimental controls, will yield accurate results.


Introduction
The understanding of global "greenhouse" issues as they relate to CO2 in the atmosphere is a current environmental concern [1,2]. To assess changes in the levels of CO2 in the atmosphere, three considerations are important: 1) a substantial history of the global levels of CO2 over an extended period of time, 2) the fidelity of standards upon which the analyses are based, and 3) accuracy and reproducibility of methods used. At the National Institute of Standards and Technology there is a continuous search for methods of analysis that yield results that are traceable to fundamental quantities. The present method used for the certification of COz/air cylinder gas mixtures as Standard Reference Materials is nondispersive infrared analysis which is calibrated with CO2 in air standards prepared by gravimetry [3]. In this work we present an analytical technique based on coulometry that provides accurate CO2 analysis for ambient-air standards having negligible quantities of other acid components. This technique can be used to assist in the validation of COJak SRMs.
In our companion paper [4] we described the design, construction, and use of an apparatus for the determination of SO2 in gas cylinder mktures at the |xmol/mol level. Here, the apparatus has been modified to demonstrate the potential utihty of coulometry as an accurate and reliable method for assessing the concentration of CO2 at the 300 to 375 jji,mol/mol level in cylinder gas mbctures.

Experimental Methods
The apparatus and experimental details have been previously described [4] and will not be discussed fully here. However, because the analytical technique is being applied to C02/air and not SO2/ N2, the experimental differences will be delineated. The experimental method is based on acid-base coulometric titration with photometric end-point detection. The experimental objectives are to: (a) select a reaction scheme; (b) develop an accurate and precise sample delivery and absorption system; and (c) determine the current efficiency independently. The absorption system selected is designed to meet the two major criteria for coulometric titration, i.e., complete absorption of the CO2 analyte gas and near 100% current efficiency [5,6,7].
The absorbing solution consists of 78 mL N,Ndimethylformamide [HCON-(CH2)2], 2 mL 0.1% thymolphthalein (5',5"-diisopropyl-2',2"-dimethylphenolphthalein) dissolved in dimethylformamide, 3 g potassium iodide dissolved in 3 mL water, and 3 mL monoethanolamine (HOCH2CH2NH2). The solution is prepared in a 100 mL flask. The pH indicator, thymolphthalein, operates over the pH range from 9.3 to 10.5 and has an absorption maximum at 598 nm. All chemicals are reagent grade and are used without further purification. A possible reaction scheme involves the reduction of H2O at the cathode.
HzO + e-^^HzH-OH- (1) and the flow rate is monitored using a calibrated mass flow controller. The relationship of cell current and flow rate of CO2 is linear over the experimental range. Equation (3) gives the relationship of cell current and flow rate as it is used to determine CO2 concentration in (.i,mol/mol. and /'|xmolCQ2^_Ccdl^A^' CO21 , .
\ mol air Q where Cceii = Cell current (A) Q = Flow rate of sample (mL/min) then MVT = Molar volume of air (corrected for temperature) mL/mol (MVT = 24475.59 @ 298 K) n = Number of electron change (e~) (n = l) Alternative schemes are presented in Ref. [5].

Procedure
The apparatus for CO2 analysis uses a narrow band interference filter (599 ±10 nm) and a titration cell, filled with the absorbing solution described above. Light passing through the clear gas-absorbing solution causes the photodiode detector to generate a titration current that is fed back to the cathode, thus causing reaction [1] to occur. The pH indicator thymolphthalein is activated by the production of the base and the absorbing solution turns blue, which decreases the output of the photodiode. When the color of the absorbing solution reaches a selected level of intensity, the feedback circuit is adjusted to give a zero current output for the cell. The C02/air mixture is introduced, metered by a mass flow controller calibrated to 0.1 mL/min. When the feedback circuit is adjusted to zero and the C02/air mixture is introduced into the cell, the analytical procedure and data analysis are identical to those discussed in Ref. [4].
The critical measurement parameters are cell current (A) and flow rate (mL/min). The current is measured using a calibrated digital current meter

Results and Discussion
Ten compressed gas cylinders with CO2 concentrations from 300 to 375 p,mol/mol were previously analyzed by a non-dispersive infrared (NDIR) analyzer calibrated with gravimetrically prepared standards [3]. A cylinder representing the mid-range of the set of 10 cylinders was selected as a control (cylinder 4, Table 1) to be used throughout this study to evaluate the overall uncertainty associated with the coulometric method and the day to day imprecision of the analysis. The control was studied over a variety of experimental conditions and for an extended period of time. The control cylinder was analyzed by the coulometric method on 17 separate occasions. Figure 1 shows the uncorrected data for the control cylinder as they were obtained chronologically over a 11 month period. The mean and the 95% confidence interval are indicated by the horizontal solid and dashed lines, respectively. The best estimate of the true mean concentration and its uncertainty at the 95% level of confidence is 346.4 ± 4.2 jjLmol/mol. For any given day, the standard deviation of a single measurement for the concentration range from 300 to 375 ijumol/mol C02/air is 2.1 |imol/mol. The reported NDIR concentration of the control cylinder was 341.7 ±0.4 ixmol/mol. The comparison of this value to the coulometrically determined value is a measure of the current effi-ciency. At the 95% level of confidence, the current efficiency is calculated to be 101.3 ± 1.0%. This value represents a systematic bias in the measurement system and gives the factor by which the experimental data should be corrected. Results for the coulometric analysis of the 10 C02/air gas cylinders mixtures as well as those obtained by NDIR are given in Table 1. The difference between the NDIR and corrected coulometric analytical values range from 0.003 to 1.1%. These results, however, show that at the present stage of development the method is capable of producing results that are comparable to those of the reference method. Figure 1 shows that the resulting analysis of the control cylinder is biased by 1.3%. No explanation for the bias is being offered at the present, however, future work is planned to improve both the precision and accuracy of the method.

Conclusion
A coulometric method for the direct analyses of reference cylinder gas mixtures has been demonstrated. It has been shown to be sensitive, potentially precise and accurate for determining CO2 in air at atmospheric concentrations, as well as a viable method of assessing the concentration of CO2