Monthly Archives: December 2022

First tests of a Fe-EDDHA|Mn-EDTA system, towards a Fe/Mn flow battery at neutral pH

18 Replies

I have recently been working on a project to create a DIY flow battery using Fe/Mn salts. The idea is to be able to achieve a close to or neutral pH system, with low cost salts, high concentrations of active species and good cycling ability. In today’s post I will describe some of my very preliminary results using a split cell system.

The image below shows you the experimental setup I am using. Both the right and left side contain graphite rod electrodes. The two chambers are separated by the DIY high permselectivity membrane I prepared using PVA/citric acid/phosphoric acid. The chamber on the left contains a solution of NaFeEDDHA from a commercial fertilizer source at a concentration of 0.05m + 3.5m of NaCl, while the cell on the right contains a solution with 0.05m of Na2MnEDTA + 3.5m NaCl. The pH was set to 7 using potassium carbonate (only a few milligrams were needed). Both chambers are stirred using magnetic stirring bars (tiny ones at 2mm).

A picture of the Fe-EDDHA|Mn-EDTA system. The left side has the Fe and the right side has the Mn. Both solutions are prepared at 0.05m concentration with 3m NaCl. The pH of the system is 7. System is showed after 2mAh of charge.

The idea of these first experiments at low concentration was to put some charge into the system to observe if there was any precipitation of Mn oxides on the cathode, or any other noticeable side reactions. We can also determine if there is any self-discharge due to crossing of Fe-EDDHA over the membrane by seeing the color change on the Mn-EDTA side and tracking the potential. I also wanted to observe what the potential was after charging (predicted standard potential is around 1.2V).

It is worth noting that the separation between the electrodes is quite large and the electrode area is low, so there are expected to be very substantial ohmic losses in this type of configuration. This means it is not useful for charge/discharge cycle data. However we should be able to get some important information about the reversibility of the chemistry and the presence of any bad side reactions, as mentioned above.

The capacity of the system at this (15mL per side) configuration would be 20.1mAh. I charged it to 2mAh at 2.3V, which was able to introduce current at a rate between 700-800mA. After stopping the charging process, the potential dropped to around 1.1V fast and then very slowly from that point. It will take more charge for the potential to hold steady there, but this already shows the chemistry is working. Changing the electrodes for new graphite rods had the potential still holding at similar values, which means the potential is not due to any deposits on the graphite electrodes.

Despite the big charging over-potential – due to ohmic losses – there was no depositing of metallic Fe on the anode or the evolution of any hydrogen gas (no bubbling was observed). I also could not observe the formation of any MnO2 precipitate on the cathode. This therefore means that the Mn3+ is stable, at least in the short term, in the catholyte (as expected from literature experienced with Mn-EDTA).

Nafion equivalent permselectivity values using a DIY PVA/Cellulose cation exchange membrane

6 Replies

During the past couple of weeks I have been working on cation exchange membranes using PVA/cellulose (see here, here, here). The idea is to create a membrane that can replace Nafion in a pH neutral flow battery built using an Fe anolyte and a Mn based catholyte. In this post I will share the first results that are up to par with those of a Nafion membrane.

My initial idea was to both crosslink and add anionic sites to the PVA by using phosphoric acid with urea as a catalyst, heating the membrane to >150C in order to perform the esterification process. This worked to a decent degree, achieving membranes with permselectivity values above 80% with sheet resistance values around 10x those of Nafion membranes.

Membranes annealed at 150C (left) and 100C (right)

However there were some obvious issues with this process. The first is that the membranes produced had some stability issues, their permselectivity would drop with time – due to lack of enough crosslinking – and the mechanical stability of the membranes also left a lot to be desired. Both of these issues were likely due to limited crosslinking of the membranes, as forming a double phosphoric acid ester is not a very favorable process, even in the presence of urea.

I would see Fe-EDDHA-1 leak across the membrane within around 24 hours of setting up my cells, with the transparent side turning a slight pink within that timeframe. The permselectivity would also drop from 80% to around 40% within that timeframe. It was obvious that these membranes had components that were still dissolving or at least creating cavities that allowed too much water to flow through.

Thinking about this, I searched for possible crosslinking agents to enhance the issue. I also wanted to avoid usage of anything toxic, like glutaraldehyde, as the space where I carry out these experiments has limited ventilation, plus I want to avoid exposing myself or my cats to harmful substances. Expensive substances were also out of the question, like sulfosuccinic acid.

Multiple results of membranes being prepared using this method (each one is around the diameter of a 25c US coin)

Reviewing papers on the subject, citric acid appeared to be a viable substance. It is a tricarboxylic acid, so it would be able to crosslink cellulose with PVA, PVA or cellulose with themselves and also keep some exposed anionic carboxylic groups to provide cation exchange capacity. Adding phosphoric acid would also catalyze the esterification reaction plus also provide some phosphorylated sites for enhanced permselectivity.

The process for preparing these membranes is as follows:

  1. Prepare a solution by adding 15g of PVA to 200mL of water (solution A).
  2. Place solution A in a fridge for 48 hours, with occasional stirring/shaking. Surprisingly, cold conditions are much better for dissolving PVA because they discourage agglomeration.
  3. Wait till solution A is fully homogeneous, keep longer in fridge and shake/stir as needed.
  4. Prepare another solution by using 0.5mL of phosphoric acid (81%), 0.5g of citric acid and 15mL of solution A. This solution is stirred until everything is completely homogeneous (solution B).
  5. Dip a filter paper in Solution B. I used Stony Lab 101 but other fine grain filter papers should work just as well. Make sure all excess has dripped off and tap with paper towels to remove any excess.
  6. Place on a hot plate at 80C for 3min
  7. Flip it to the other side for another 3 minute.
  8. Use a brush to paint solution B on the filter paper while on the hot place.
  9. Wait for 3 minutes.
  10. Flip the filter paper and paint the other side, wait another 3 minutes.
  11. Repeat steps 8-10 three times.
  12. Increase the temperature to 150C.
  13. Flip the membrane every 10 minutes for one hour or until the membranes appear fully black. Put a petri dish on top if needed to keep the membrane flat.
  14. Allow the membrane to cool to room temperature.
  15. Place the membrane in a solution with 10g/L of potassium or sodium carbonate to neutralize any remaining acid, they can be stored in a 0.5M NaCl solution.

The membranes that result from this process are black in nature. However they do not feel like charcoal and do not crumb easily. Instead, they have the feeling of a piece of plastic film, which is exactly what we are looking for. Several papers discussing citric acid crosslinking of different polymers do have resulting black films, so this isn’t necessarily a bad thing.

I was also very pleasantly surprised by the permselectivity measurement for these membranes. Measuring the potential across NaCl 0.1M | NaCl 0.5M using identical Ag/AgCl reference electrodes, the potential is 38-39mV, meaning that these membranes have permselectivity values >99%, which is equivalent to that of the best Nafion membranes. Adding 0.01g of NaFeEDDHA to the NaCl 0.5M side – which makes it dark red – I could see absolutely no crossover of FeEDDHA-1 to the other side of the half-cell experiment within 48 hours of testing. There were also no drops in the permselectivity which remains extremely high. The sheet resistance measurements are also very favorable, with in place sheet conductivity values now in the <50 ohm/cm2 range.

Overall, I am pleased with this DIY membrane result. The crosslinking of PVA using citric acid and phosphoric acid on a cellulose matrix provides you with a very robust membrane that has some wonderful characteristics. This will be my base membrane for the construction of Fe-Mn flow batteries. This membrane is also very low cost.

Measuring and improving the performance of PVA/Cellulose cation exchange membranes

2 Replies

In a previous post I described how to create a DIY cation exchange membrane using some easy to get materials. These membranes could achieve significant permselectivity values, but still far away from those required to create membranes for a robust flow battery. Additionally, the sheet resistance of these membranes – which I measured using a 4 contact electrode method – was quite bad, with values often greater than 6000 ohm/cm2. The through plane resistance was around 3x that, although my method for through-plane resistance measurement is not reliable yet.

Some of the last membranes I produced using a PVA solution with a pH in the 6-7 range. The membrane remains an off-white yellowish color, but does not oxidize as in my previous tests.

In this post, I want to talk about the advancements I have made to improve the fabrication of these membranes. First of all, I have lowered the preparation temperature to 150C, this avoids charring the membranes and improves reproducibility. I also added 80 minutes of additional time at these temperature once all the PVA coats have been put on, this improves crosslinking and drastically reduces the solubility of the membrane in water (to the point where it becomes fully insoluble).

I have also found out that decreasing the acidity by adding some potassium hydroxide also helps retain membranes structure, increase permselectivty and decrease sheet resistance. This matches some papers on cellulose phosphorylation using potassium phosphate and ammonium phosphate salts, with solutions that have much higher pH values than phosphoric acid. The higher pH helps preserve the structure of the cellulose and PVA, as a lack of acid reduces the changes of degradation of the cellulose and PVA. The phosphorylation still happens, thanks to the urea catalyst present.

With this in mind, the membranes can probably be made using monopotassium or monoammonium phosphates, much more readily and less dangerous chemicals compared to concentrated phosphoric acid and potassium hydroxide.

One of my last experiments to measure permselectivity. The cell to the right contains a very small amount of NaFe(EDDHA), which has a very deep red color. This makes it very easy to see membrane crossover.

The best values I have achieved so far are a permselectivity of 80% and a sheet resistance of 373 ohm/cm2. These are still much worse than those of commercially available membranes, but certainly better than the values I was achieving before.

From the parameters I have tested, the cross-linking temperature and pH seem to be the most important to the qualities of the membrane, so I will try to study these too with a bit more detail to find out if I can produce membranes with better qualities. Increasing the concentration of P at higher pH values with higher urea quantities might also help achieve better cross-linking.