Vol.2 No.3 2009

Research paper : A marked improvement in the reliability of the measurement of trace moisture in gases (H. Abe)−209−Synthesiology - English edition Vol.2 No.3 (2009) from ice decreases with the decrease in temperature, and it becomes more sensitive to external thermal stimuli as the vapor pressure becomes more susceptible to temperature change. Moreover, even when saturation is achieved, it is not easy to confirm the state of equilibrium, and therefore, a large degree of uncertainty must be taken into account. These issues can be avoided in the diffusion tube method. For determining the amount-of-substance fraction of water in the generated gas, the frost point method requires the vapor pressure formula and enhancement factor. In contrast, the diffusion tube method does not suffer from this problem, because this method can be used to determine the amount-of-substance fraction of water in the gas from the measurements of the evaporation rate and flow rate, and therefore, it is a method clearly traceable to the SI to generate a humid gas with a known amount-of-substance fraction of water. Because Sonntag’s equation,[7][8] which is often used as the vapor pressure formula for ice, is only effective down to -100 ºC, the amount-of-substance fraction of water cannot be determined when Sonntag’s equation is used for temperatures below -100 ºC in the frost point method. Moreover, there have been no reports of experiments using the enhancement factor below -35 ºC, and it has been pointed out that the extrapolation of the formula is unreliable below this temperature[9]. This is not an issue for the diffusion tube method, which neither requires the vapor pressure formula nor the enhancement factor.However, there are disadvantages in the diffusion tube method, one of which is the need to measure a very small evaporation rate. For the generation of trace moisture of 14 nmol/mol (ppb) using the diffusion tube method, the evaporation rate becomes extremely small, approximately 14 µg/h (780 nmol/h) when the dry gas is nitrogen and the flow rate is 20 L/min (at 0 ºC and 101.325 kPa); the key is how this evaporation rate can be measured with as small uncertainty as possible. There is also the problem of zero gas. In the diffusion tube method, the moisture remaining in the dry gas may generate uncertainty, and it is necessary to prepare a dry gas called zero gas where the moisture content is lowered to as close to zero as possible. Another problem is that there have been no previous reports in which the diffusion tube method was established as a primary measurement standard of trace moisture. This meant that little of the knowledge, experience, or technology needed for the standard establishment has been accumulated, and it was necessary to design the instrument and develop the analysis method of the uncertainty almost from scratch.After carefully reviewing the advantages and disadvantages of the diffusion tube and frost point methods, AIST selected the diffusion tube method, as already mentioned. There were four primary reasons for selecting this method, as follows. First, the realization and confirmation of saturation required in the frost point method were considered problematic. Investigation of the development of the trace moisture standard started in around 1999 at AIST (formerly, National Research Laboratory of Metrology, Agency of Industrial Science and Technology), but some national metrology institutes in other countries had already started developing the trace moisture standard on the basis of the frost point method. However, to the best of our knowledge, there was no published research at that time that reported the uncertainty of the trace moisture standard around the frost point of -100 ºC on the basis of measurement traceable to the SI and that described the analysis method of uncertainty evaluation. We considered that this was attributable to the difficulty in the realization and confirmation of perfect saturation around -100 ºC. Second, the direction of future trace moisture measurement was considered. According to the International Technology Roadmap for Semiconductors 2000[10], the level at which trace moisture must be controlled had already been reported to be equivalent to the frost point of -100 ºC or less, and it stated that controlling residual moisture in the lower frost point range will be required in the future. We considered that the reliability of the vapor pressure formula and enhancement factor would become issues when the frost point method is used in this range. The third reason was the development of a new balance called the magnetic suspension balance[11]. This balance became commercially available in the early 1990s, and many research studies using this balance were published from the middle to late 1990s. By 2000, it was demonstrated that this balance has excellent performance and usability, and we considered that the difficulty in measuring the very small evaporation rate in the diffusion tube method could be overcome through the use of the magnetic suspension balance. Finally, as metrologists, we find the clear traceability to the SI of the diffusion tube method very fascinating.3.2 Development of the primary measurement standardThe amount-of-substance fraction of the trace moisture in the gas generated by the trace moisture generator using the diffusion tube method is given as (1)wherexwxbFFxb〜〜NbNNNFFNbNb+=+++++Amount-of-substance fraction of trace moisture in generated gas [mol/mol]Amount-of-substance of moisture generated as vapor from diffusion tube per unit time [mol/h]Amount-of-substance of moisture that moves by adsorption/desorption per unit time [mol/h] Flow rate of dry gas per unit time [mol/h]Amount-of-substance fraction of residual moisture in dry gas (zero gas) [mol/mol]xwNNbFxb :: :::


page 36