Measuring mass flow in “wet hydrogen”: what does water do to your accuracy?
Hydrogen from electrolysers, reformers or storage systems is rarely dry. Even small amounts of water vapour change the thermal properties of the gas, with direct consequences for the accuracy of a thermal mass flow controller (MFC).
⏱ approx. 5 minutes reading time
How a thermal MFC measures
A thermal MFC works on the basis of heat transfer. In the sensor, a small portion of the gas is passed along a heated element. The temperature difference before and after this element is proportional to the mass flow, but only if the thermal properties of the gas are known and constant.
The key property is the specific heat at constant pressure: Cp. The MFC is factory-calibrated on dry hydrogen with a fixed Cp value. As soon as that Cp changes due to the presence of another gas component, the meter reads systematically incorrectly.
Operating principle: Q = ṁ × Cp × ΔT → ṁ = Q / (Cp × ΔT) If Cp changes while Q and ΔT remain the same, the MFC calculates an incorrect mass flow.
The problem: water vapour has a very different Cp from H₂
This is the heart of the problem. Hydrogen has an extremely high specific heat, the highest of all gases. Water vapour has a considerably lower Cp. When the two mix, the effective Cp of the mixture drops.
The difference is dramatic: the Cp of water vapour is a factor of ~7.5 lower than that of hydrogen. This means that even a small percentage of water vapour causes a noticeable reduction in the effective Cp and thus introduces a measurement error.
When is this a problem in practice?
- Electrolysis hydrogen: hydrogen straight from a PEM or alkaline electrolyser is saturated with water vapour at process temperature. Without drying, the moisture content can rise to 2–5 vol%, depending on pressure and temperature.
- Hydrogen refuelling stations and pipelines: condensation can occur during pressure changes (filling/discharging of tanks). The effective moisture content in the gas phase varies with the process conditions.
- Fuel cells (PEMFC): for optimal membrane performance, the incoming H₂ is deliberately humidified. The MFC upstream of the cell therefore works with wet gas unless this is specifically corrected for.
- Laboratory applications and R&D: when testing H₂ components under realistic conditions, humid gas is often used deliberately. A standard MFC calibrated on dry H₂ then gives a distorted picture.
Additional effect: adsorption on the sensor surface
In addition to the change in Cp, moisture has a second, more insidious effect: water adsorption on the sensor surface. The heated sensor element and the surrounding capillary act as condensation nuclei at varying temperatures.
A thin film of moisture on the sensor element changes the heat transfer coefficient, slows down the response time of the sensor and, over repeated cycles, can lead to calibration drift. This cannot be corrected via software; it requires physical drying of the gas or cleaning of the sensor.
Solutions and recommendations
Solution 1: Dry the gas upstream of the MFC.
The most robust approach. Use a Nafion dryer (e.g. Perma Pure) or a molecular sieve directly upstream of the MFC inlet. Dry H₂ (< 10 ppm H₂O) gives optimal MFC accuracy.
Solution 2: MFC with Cp correction for gas mixtures.
Some MFCs support a gas mixture mode in which you can enter the Cp of the actual mixture. For a known and stable moisture concentration, this is a workable approach.
