Testing Gas Sensors Over a Range of Relative Humidities
with applications to the generation of ternary gas mixtures.
For any sensor that will be used to monitor the environment, the effect of relative humidity (R.H.) must be measured. Many gas sensors are upset by changes in relative humidity (R.H.). Certain sensors, such as conductive polymers and solid-state (tin oxide) types, are notorious for their R.H. effects. Electrochemical sensors are sometimes touted as unresponsive to R.H., but this depends on both the sensor and the gas. In certain gold-electrode sensors, H2S may have a strong humidity dependence, and sulfur dioxide will not.
Measuring R.H. coefficients is a common example of a ternary gas-mixing problem: you want to mix the test gas, the carrier (usually air or nitrogen), and water vapor, so that all three come out to predetermined concentrations. Ternary gas mixtures are very tedious to make by the usual methods, and prone to error. With one or two Model 1010 Precision Gas Diluters and a hand calculator, you can make these mixtures in a short time, and depend on their accuracy.
Lesson One in making R.H.-controlled mixtures: Don't use bubblers to make your water-saturated gas. A simple bubbler is unlikely to achieve an R.H. above 70%. Even using fine spargers and deep water columns, it is hard to exceed 85% R.H. Furthermore, the bulk water will cool as evaporation proceeds, and the R.H. will gradually drift downward over time.
The conventional method of making water-saturated air is to boil water in a stream of air, and pass the steam through one or two condensers actively held at room temperature [Nelson, 1992]. This is a great pain if you only plan to make a few measurements. You can fill a large sample bag with 100% R.H. air by injecting a measured amount of liquid water into a bag filled with a known volume of dry air, and evaporating it with a heat gun. If you use a sample bag of a transparent material like TedlarTM, you can see the water drops. This simplifies the process considerably. After the humidified air has cooled to room temperature, it is ready for use. (Don't forget to start with dry air from a cylinder, or excess water will condense in the bag!)
Now to set up for our tests: We have a cylinder of dry air and a cylinder of standard ethane in air at 5,000 ppm. We want to test our sensor at 0, 100, 300, and 1000 PPM ethane, at 25%, 50%, and 75% R.H.
The schematic for the test system is shown in Figure 1. In this case, two Model 1010-P diluters are used, so that balancing adjustments can be made to keep one of the constituents constant while varying the other. Notice that a vented bottle is located between the two diluters. The function of the first diluter is to keep the vented bottle filled with air of controlled humidity; the second diluter will draw from this bottle as needed. Excess humidified air is vented from the top of the bottle.
