Filtern
Dokumenttyp
- Beitrag zu einem Tagungsband (5)
- Zeitschriftenartikel (3)
- Posterpräsentation (3)
- Vortrag (1)
Sprache
- Englisch (12) (entfernen)
Schlagworte
- Coulometric sensors (7)
- Trace humidity measurement (6)
- ANOVA (2)
- Chemical reaction (2)
- Coulometric sensor (2)
- Gas pressure influence (2)
- Measurement uncertainty (2)
- Response behavior (2)
- Sensor coating (2)
- Temperature influence (2)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (1)
A portable device for calibration of trace humidity sensors and an adopted calibration procedure have been developed. The calibration device is based on humidity generation by permeating water through polymeric membrane tubes. Water vapour transmission rates for various polymers were experimentally determined in order to select the most suitable polymeric material. The developed trace humidity generator consists of a gas-flow polymeric hose immersed in a water reservoir thermostated by a sensorcontrolled heater. Mole fractions of water vapour between 1 µmol/ molˉ¹ and 350 µmol/molˉ¹ (equivalent to frost-point temperatures from -76 °C to -31 °C) were generated by varying either the operating temperature or gas flow. The operating temperature can be varied from 20 °C to 60 °C and kept stable within 0.1 K. Uncertainty analysis indicated that the trace humidity generator produces gas flows of constant humidity amounts with a relative expanded uncertainty less than 3.4% (k = 2) of the generated value.
Coulometric sensors are applied for trace humidity measurements in various technical gases. The use of this sensor type is demanded in some standards by the European Pharmacopoeia for medical gases. Coulometric sensors allow the measurement of water vapour in gases such as e.g. air, Cl2, H2, N2, N2O, CH4 between the concentration ranging from 0.1 to 2,500 µmol∙mol-1 which corresponds to frost point temperature -90 °C to -10 °C, respectively.
The sensing principle is based on Faraday’s law of electrolysis whereby water is decomposed to hydrogen and oxygen. The sensor signal is the measured electrical current which is proportional to the mass of water that is absorbed on the hygroscopic phosphorous pentoxide layer. The signal is dependent on the gas flow at a given voltage, gas pressure and temperature.
The sensors need to be calibrated to measure the accurate signal for humidity in air. However, the signal is dependent on the type of gas matrix. This dependency has not been quantified so far. Therefore, the impact of reactive gases such as hydrogen and nitrous oxide on the calibration curve was investigated. Furthermore, a possible rationale in relation to the interaction of water with the gas matrix and its impact on the electrode reactions is suggested.
The experimental setup consists of a gas supply, dryer, humidifier, test chamber and reference hygrometer. The test gas is generated by mixing the dry and the wet gas flow. First the carrier gas is dried by an activated carbon filter and then split into two flows. One flow is dried again with a molecular sieve. The other flow is humidified by passing it through a bubbler filled with pure water. After this, the rate of both flows is controlled by mass flow controllers and then mixed with the test gas. A calibrated precision chilled mirror dew-point hygrometer is used as a reference instrument.
The coulometric sensors were exposed to different humidified gases and the sensor signal was recorded till a constant value was obtained. Calibration curves were calculated for the frost point temperature in the range of about -70 °C to -10 °C according to the equation, tf = A + B∙ln(I), (I is the electrolysis current and A, B are constants) followed with a linear regression fit.
Comparison of the results of air and nitrogen showed no significant differences. In contrast, there are remarkable differences for humidified hydrogen and nitrous oxide, respectively.
The difference might be due to increased recombination of hydrogen with the produced oxygen to form new water molecules in humidified hydrogen.
In conclusion, coulometric trace humidity sensor is a robust hygrometer for various technical applications. However it needs to be calibrated for the specific gas matrix.
This work presents a first attempt to use ionic liquids as a new coating for planar coulometric sensors. These sensors are used for the measurement of trace humidity in various gases. Usually, the coating of the sensors is tetraphosphorus decaoxide and its hydrolysis products. Instead, a hygroscopic ionic liquid was used as sensor coating in this work. Generated frost point temperatures tf in the gas ranged from -80 °C to -30 °C, which is equivalent to vapour mole fractions xv from 0.5 μmol·mol-1 to 376 μmol·mol-1. In addition to the coulometric sensors, the generated humidity is determined by a precision dew point hygrometer as reference. First results show that it is possible to measure humidity withionic liquid based coulometric sensors.
This work presents a first attempt to use ionic liquids as a new coating for planar coulometric sensors. These sensors are used for the measurement of trace humidity in various gases. Usually, the coating of the sensors is tetraphosphorus decaoxide and its hydrolysis products. Instead, a hygroscopic ionic liquid was used as sensor coating in this work. Generated frost point temperatures tf in the gas ranged from -80 °C to -30 °C, which is equivalent to vapour mole fractions xv from 0.5 μmol·mol-1 to 376 μmol·mol-1. In addition to the coulometric sensors, the generated humidity is determined by a precision dew point hygrometer as reference. First results show that it is possible to measure humidity with ionic liquid based coulometric sensors.
Planar coulometric sensors were investigated in humidified synthetic air at various absolute gas pressures, i. e. 2 bar, 5, bar, 10 bar, and 15 bar. Humidified gas flow at adjusted gas pressure was split into two flows, one passed a coulometric sensor and the other one passed a reference hygrometer after decompression. Both signals were recorded and then compared after calculation of resulting frost point temperature. Calculation is based on a calibration function obtained at ambient pressure. Comparison showed that an increasing pressure resulted in a higher derivation between sensor signal (calculated frost point temperature) and reference frost point temperature. At an absolute pressure of 2 bar the differences were minor in consideration of the uncertainty, however at 15 bar the differences were 6.77 K. Nevertheless, it was possible to measure the gas humidity at higher pressure with coulometric trace humidity sensors.
Planar coulometric sensors were investigated in humidified synthetic air at various absolute gas pressures, i. e. 2 bar, 5, bar, 10 bar, and 15 bar. Humidified gas flow at adjusted gas pressure was split into two flows, one passed a coulometric sensor and the other one passed a reference hygrometer after decompression. Both signals were recorded and then compared after calculation of resulting frost point temperature. Calculation is based on a calibration function obtained at ambient pressure. Comparison showed that an increasing pressure resulted in a higher derivation between sensor signal (calculated frost point temperature) and reference frost point temperature. At an absolute pressure of 2 bar the differences were minor in consideration of the uncertainty, however at 15 bar the differences were 6.77 K. Nevertheless, it was possible to measure the gas humidity at higher pressure with coulometric trace humidity sensors.
Planar coulometric sensors were tested in humidified synthetic air at various gas temperatures. Generated frost point temperature in the gas ranged from -30 °C to -60 °C and were measured by coulometric sensors and in addition by a calibrated dew point hygrometer. The gas temperatures, which were measured by a calibrated Pt100 sensor, were set to -20 °C, 0 °C, 23 °C, 40 °C, 50 °C, and 60 °C during the experiments. Empiric nonlinear functions were calculated between the generated humidity and the sensor signals. In comparison to the measured signals at 23 °C, the sensor signals were lower at the other gas temperatures. The measurements at 60 °C showed indistinct results due to a great signal noise. The response behavior of the sensors was similar at 23 °C, 40 °C and 50 °C. In contrast to that, the sensors reacted slowly at a gas temperature of -20 °C and 0 °C. In summary, with coulometric sensors it was possible to measure continuously trace humidity with an expanded uncertainty below 2 K.
Coulometric sensors were tested in humidified synthetic air at various gas temperatures. Generated frost point temperature in the gas ranged from -30 °C to -60 °C and were measured by coulometric sensors and in addition by a calibrated dew point hygrometer. The gas temperatures, which were measured by a calibrated Pt100 sensor, were set to -20 °C, 0 °C, 23 °C, 40 °C, 50 °C, and 60 °C during the experiments. Empiric nonlinear functions were calculated between the humidity and the sensor signal. In comparison to the measured signals at 23 °C, the sensor signals were lower at the other gas temperatures. Measurements at 60 °C showed indistinct results due to a great signal noise. The response behavior of the sensors was similar at 23 °C, 40 °C and 50 °C. In contrast to that, the sensors reacted slowly at a gas temperature of -20 °C and 0 °C. In summary, with coulometric sensors it was possible to measure continuously trace humidity with an expanded uncertainty below 2 K.
Trace humidity was measured by using miniaturized planar coulometric sensors in technical gases such as hydrogen, nitrogen, helium, nitrous oxide, and synthetic air. Frost point temperatures tf in the gases ranged from −60 °C to −30 °C, which is equivalent to a vapour mole fraction xv from 10 μmol mol−1 to 376 μmol mol−1. In addition, the generated humidity was determined by using a precision dew point hygrometer as reference. Nonlinear calibration functions were calculated that correlated electric current (sensor signal) and reference humidity. Parameters of functions were tested with one-way analysis of variances (ANOVA) to prove if all used sensors had a similar behavior in the same gas during experiments. Results of ANOVA confirmed that averaged functions can be applied for trace humidity measurement in nitrogen, helium, nitrous oxide, and synthetic air. The calculated functions were negligibly different for nitrogen, helium, and synthetic air. In humidified nitrous oxide, a minor change of parameters was observed due to lower electrical currents. In total contrast to that, the measured sensor signals were significantly higher in humidified hydrogen and each sensor required its own calibration function. The reason was a recombination effect that favoured multiple measurements of water molecules. Nevertheless, it was possible to measure continuously trace humidity in all tested gases by using coulometric sensors with an expanded uncertainty below 2 K (k = 2).
Especially trace amounts of water vapour in gases can be reliably determined by coulometric trace humidity sensors. The principle of these sensors is based on water vapour absorption in a hygroscopic layer and its subsequent electrolytic decomposition. The calibration of sensors was performed in the humidity range, expressed as frost point temperature, from −30°C to −80°C . This range is equivalent to volume fractions smaller than 376 µL·L−1. Generated humidity was measured with coulometric sensors and a chilled dew point hygrometer that was used as reference. An empirical non-linear function was found between sensor signal and measured reference humidity. This function consists of two parameters with a measurement uncertainty. Both calibration parameters were checked by means of one-way analysis of variance. It showed that gas specific function can be used for humidity measurement in nitrogen, hydrogen, dinitrogen monoxide, compressed and synthetic air. It is possible to determine trace humidity in all tested gases with an expanded uncertainty less than 2.1 K (coverage factor k=2 ) regarding frost point temperature.