Q&A with HYDErogen part 4
Under the spotlight: Hydrogen electrolyser technology part four
Continuing our interview with guest contributor Dr Kris Hyde, we spoke to him about the basics of hydrogen electrolyser technology and the challenges faced from the various options available. This time, the focus is on the influence of system water.
Dr Hyde, a specialist in electrolysis technology, is director of HYDErogen, a consultancy firm offering technical and commercial advice on hydrogen-related technology and business.
What is electro-osmotic drag and how does it influence PEM electrolyser design?
In a PEM electrolyser, water is split on the anode into oxygen and two protons, which travel through the membrane to the cathode, where they combine to form hydrogen. However, being charged, the proton attracts polar water molecules, such that every proton transversing the membrane, drags across ~four water molecules. So, for every water molecule destroyed in the electrolysis process, eight are lost to electro-osmotic (EO) drag. If that EO water was not recycled, the electrolyser water consumption will be an order of magnitude higher.
A 1MW electrolyser will transfer ~30t/day of EO water/day (depending on the electrolyser efficiency: an efficient low cell voltage needs a higher current to consume the same power. Higher current leads to higher EO drag). Thus, EO drag leads to PEM electrolysers usually requiring significant additional equipment often including water separator filters, chillers and separation vessel. Fortunately, PEM electrolysers usually operate at differential pressure between H2 and O2. So, one can connect a pipe to the bottom of the separation vessel, and when it’s full, open a valve and the H2 pressure will push the water round to the anode for recycling. However, this should be done with caution:
Henry’s law states that the concentration of hydrogen in the EO water is proportional to the pressure. In moving from 30 bar (typical of PEM electrolysers) to atmospheric pressure, ~10L/min of H2 is released for a 1MW electrolyser. Rather than deliberately dump this H2 into the O2 side, it is often extracted and vented (and represents a very small drop in the systems faradaic efficiency). However, even once depressurised to atmospheric, some H2 will remain dissolved, which can cause unwanted side reactions on the anode.
The anode water in a PEM electrolyser is kept very pure (typically between one and 10 MΩ-cm), but EO drag water has high levels of impurity. Thus, engineers purify this water before it enters the O2 water loop. This could involve feeding into the potable water purification system, or the anode water polishing loop if present.
The control system will see this water transfer as a drop in H2 pressure, particularly if operating at min operating point. This could be interpreted as an H2 leak and trigger an alarm; thus, the sensitivity of this safety alarm cannot be as high as desired.
Image courtesy of HYDErogen
What if the separation vessel level sensor or exit valve fails such that once the EO water has been pushed across, the H2 follows?
This could potentially lead to an explosive atmosphere in the incoming water purification system or the anode separation vessel. Thus, it is common to see independent water level monitoring on the H2/Water separation vessel.
How much water does a 1MW PEM electrolyser stack need?
So, let’s start with some typical numbers. A 1MW PEM stack with 200 cells running at 3A/cm2 @ 2.0V, gives a cell area of 833cm2, so a diameter of 33cm. I have assumed circular stack as PEM systems are generally pressurised and circular is the best shape to constrain. Some basic chemistry tells you that the process of electrolysis coverts 0.23ml of water per sec per cell to hydrogen. However, following the electo-osmotic drag analysis above, we’re up to 2.1 ml per sec per cell.
However, each cell (based on HHV) is releasing 1.7kW of heat, so we need water going into the cell to remove the heat. The amount of water depends on how much temperature variation across the cell is acceptable. For now, let’s limit it to 4°C difference between input and output water temp. Assuming even heat removal, this requires 78ml per sec per cell. However, in the absence of a flow field (which are expensive) circular cells have very uneven heat removal. Imagine having an ‘in’ port on one side of the cell and an ‘out’ port on the other. Flow will be straight across the middle with large amounts of dead space at the sides with no flow. We can try various things to improve this (multiple, angled ports, baffles etc), but ultimately to ensure the edges are limited to a 4°C rise, you need far more flow than 78 ml per sec per cell. Design depending, with an additional factor of five is not unreasonable. This brings us to about 388ml per sec per cell, giving a total flow of 78L/sec to the 200 cell stack.
So how much of the water flow into the PEM stack goes to electrolysis? About 0.06% and the rest of the water is pumped back around the system.
***
More about HYDErogen
HYDErogen can offer technical due diligence services, help solve real-world problems and is backed by an engineering sister-company to produce original equipment. Contact Kris Hyde on LinkedIn.