Validation of the NPL gravimetric hygrometer

The gravimetric hygrometer of the UK National Physical Laboratory (NPL) provides a first-principles realisation of mixing ratio - the most fundamental measure of the humidity of a gas. The operation of the NPL gravimetric hygrometer has been validated to demonstrate its satisfactory performance as a primary standard for humidity within the UK national measurement system. The measurement performance of the gravimetric hygrometer has been characterised for mixing ratios within the range 0.01 g kg to 155 g kg 4 (equivalent at atmospheric pressure to a dew-point range of -60 °C to +60 °C). The significant aspects of the measurement have been assessed in detail; these being the balances and weighing processes, the efficiency of collecting water and dry gas, and the effects of stray water in the instrument. Sources of systematic error have been identified and their effects quantified. Wherever possible, such errors have been eliminated or reduced. Elsewhere, numerical corrections have been evaluated and applied to the results of measurements. The overall uncertainty of measurement for the gravimetric hygrometer has been estimated by evaluating the uncertainties contributed by each aspect of the measurement, and combining these statistically to find the overall effect. The estimated uncertainty at a level of confidence of approximately 95% (a coverage factor of k=2) was found to range between 0.015 percent of value at the highest humidity considered, near 155 g kg 1, and 1.27 percent of value at the lower limit of 0.01 g kg1. The gravimetric hygrometer has been intercompared with the humidity generator which is the NFL standard for dew point, using nitrogen as the carrier gas in these measurements. For the conversion between values of dew point and mixing ratio, the uncertainties in the reference functions have been reviewed for the vapour pressure curve of water, and for the water vapour enhancement factor which accounts for the non-ideal behaviour of humid gases. A new calculation of the enhancement factor for nitrogen is presented. The uncertainties due to sampling were also evaluated. On average, the uncertainties due to the reference functions and those due to sampling were found to comparable with the uncertainty in the gravimetric measurements. Taking all these into account, the intercomparison measurements agreed to within the combined uncertainties of the two instruments.