A Simple Analyser for Gases In Mixed Samples

Several simple analysers have been described that perform such an analysis by employing an indexed chamber of fixed size to measure the volume of a sample (Scholander & Roughton, 1942; Berg, 1946; Scholander, Van Dam, Lloyd Claff & Kanwisher, 1955). This paper describes a simple analyser which employs a chamber of variable size to measure the volume of a sample. As a result of its design this instrument accepts a larger range of sample sizes and demonstrates higher accuracy in volume determination (and hence overall accuracy) than previously described instruments.

The analyser is drawn in Fig. 1. In operation the glass tube is filled with approximately 400 μl of an inert solution, after which 50–150 μl of sample is placed inside the outer reservoir (OR) with a syringe. The bubble of sample gas is then drawn through the capillary towards the inner reservoir (IR) by using the micrometer to increase the volume of the IR. Readings are taken from the micrometer as the bubble first and then last touches the reference mark on the capillary as it is pulled towards the IR. The difference between the two readings indicates the volume of the sample. Once the volume of the sample has been determined it is moved completely into the IR and the inert solution in the OR is replaced with approximately 300 μl of a reagent which absorbs the gas of interest. The volume of the IR is then increased to draw the new reagent through the capillary into the IR where it encounters the sample. The sample is then moved back to the OR by decreasing the volume of the IR. The several seconds this movement takes are sufficient for the new reagent to absorb completely the gas of interest. The sample is then remeasured as above and the change in volume is compared with the original volume of the sample, giving the concentration of the absorbed gas. Following this,the sample may again be moved completely into the IR and the process repeated to determine the concentration of another constituent of the sample.

The results of the analyses of three samples of N₂ containing CO₂ and O₂, in physiological ranges, and ambient air are given in Table 1. All samples were approximately 100 μl in volume. In each analysis the volume of the sample was measured first in acid-citrate to determine total volume, second in KOH to absorb CO₂, and third in pyrogallol to absorb O₂ (all solutions as per Scholander et al. 1955). The gas concentrations in each sample were calculated as:

The average of the standard deviations across all samples indicates that the precision of the analyser, as ±1 S.D., is approximately ±0·2%. The average of the discrepancies for each gas across all samples indicates an error in accuracy of – 0·11 % for CO₂ and –0·23 % for O₂. These two discrepancies are representative of only three samples each, taken from a reference standard which varied by ±0·015 %. Therefore they may not be representative of the true accuracy of the instrument.

Nonetheless, these latter averages indicate the presence of a small systematic error in the determination of the percentage concentration of CO₂ and O₂ by this analyser.

With this instrument, the size of a sample is directly related to the precision with which its volume may be measured. The micrometer head is advertised by the manufacturer as having an accuracy in delivery of ±0·2 μl, so a 100 μl sample should show a maximum precision of ±0·2% (as shown by the data), a 50 μl sample a maximum of ±0·4%, a 25 μl sample a maximum of ±0·8%, and so on. However, this instrument does not constrain the use of a particular sample size within a 1-150 μl range in order to resolve ±0·2 μl (cf. the maximum volumes of Scholander et al. 1955, of 28·3 μl or Berg, 1946, of 0·46 μl). With the substitution of a more accurate micrometer head, the precision of the instrument would increase.

The factor capable of producing the greatest error is temperature change. A change of 0·1 °C will cause a volume change of just under 0·4gl in a 100 μl sample. The water jacket fitted to the instrument makes this error small. CO₂ is very soluble in water, so a high-salt acidic formula is used for the initial solution in order to prevent CO₂ dissolution. Although no inaccuracy due to CO₂ solubility is shown in the data, samples with CO₂ concentrations of 15 % or more could be expected to show some loss into the initial solution in spite of its composition. Oxygen is much less soluble by comparison and should show little solubility at higher concentrations. There is some problem predicting these solubilities because, although calculations will show that a significant proportion of a sample should go into solution, this does not occur in practice. It is thought that there is insufficient mixing during an analysis to produce the expected amount of gas dissolution.

The accuracy and simplicity of use of this instrument, coupled with its ability to analyse for O₂ and CO₂ in the same sample, may make it suitable for instructional as well as research applications. This instrument is available from the author in limited quantity.