An alternative approach to CO2?

[Yugang]

We do not need experiments to identify what controls the diffusion rate here; Fick’s law already tells us the relevant parameters.

The problem with Fick's law is that we don't know the Diffusion coefficient of CO2 in water. The often quoted 1.8×10−9 may or may not apply to our typical aquarium conditions considering turbulence and surface agitation. But I agree, we could use it at least as a first guestimation, as starting point.
Then of course, we would need an estimation for the plant CO2 uptake, in relation to the CO2 diffusion in water, as these two are the only relevant CO2 flows in steady state.

Maybe you could also inject 100% CO2 into the water instead of the headspace to have more efficient diffusion into the water (exploiting a higher pressure difference between the bubbles and water, and a larger surface for efficient delivery of CO2), and still measure the CO2 in the headspace.

Using the sensor with the "reverse" setup, ie injecting in the water while measuring above, is essentially how professional CO2 meters work and could save a couple of thousand USD. The benefit will be that we take the water chemistry out of the equation, and could be much better than a pH probe for controlling CO2.

air pump and diffuser that takes air from the headspace and injects it into the water, therefore increasing the surface for gas exchange

This seems not a very good idea to me. Air pumps for our hobby don't have much pressure, the bubbles will therefore be quite large and hence the surface area of bubbles will not be very large compared to the surface area of the tank's water. For sure it will help, but probably not too much.

 

[hax47]

The problem with Fick's law is that we don't know the Diffusion coefficient of CO2 in water. The often quoted 1.8×10−9 may or may not apply to our typical aquarium conditions considering turbulence and surface agitation. But I agree, we could use it at least as a first guestimation, as starting point.
Then of course, we would need an estimation for the plant CO2 uptake, in relation to the CO2 diffusion in water, as these two are the only relevant CO2 flows in steady state.

What I was trying to refer to in Fick's law is that the diffusion rate is directly proportional to the partial pressure difference. And there we have more than >=200x difference when we compare 1% CO2 to 100% CO2. The exact surface of the agitated water is a more complicated problem, and the diffusion coefficient could be problematic, too. But that is irrelevant if we compare the two CO2 pressures. That is also why I used a modified form of Fick's law before, introducing diffusion capacity (or we can call it conductance or whatever), combining the surface and the coefficient into one component. This can then be determined in a simple experiment (following pH over time after injection of CO2) for each aquarium setup.

This seems not a very good idea to me. Air pumps for our hobby don't have much pressure, the bubbles will therefore be quite large and hence the surface area of bubbles will not be very large compared to the surface area of the tank's water. For sure it will help, but probably not too much

I addressed this exact question in the previously linked experiment, and the air pump was more effective than the strong surface agitation (check the gifs for the agitation level). Compared to a "normal" agitation, the effectiveness of air pumps could be several-fold better.

What I would also conclude from these plots, at least with the plant density in my aquariums, is that plants were more effective in removing CO2 from the water as the diffusion through the surface at similar pressure difference ranges you plan (compare the slope of the CO2 decrease between the lighting and dark periods). So the diffusion rate between air and water was less than the rate of plant consumption.

Don't get me wrong, I am not trying to be dismissive at all — I like this idea, I just think it is worth considering these points in advance so that the system can be made as effective as possible.

 

[PeerUnk]

@Yugang @hax47 great discussion guys! I already planned to used a small fan for the sensors setup in order to increase speedy and more accurate measurements. It will help the diffusion rate as well.

Ofcourse water movement should be incorporated as well in the test.

I will try to install the pH sensor device as well, so we can track diffusion rates of the CO2.

If the diffusion rate is too low in practice, I will use a small piezoelectric air pump (I own this one). I assume it will help big time.

Cheers!

 

[Yugang]

Don't get me wrong, I am not trying to be dismissive at all — I like this idea, I just think it is worth considering these points in advance so that the system can be made as effective as possible.

Not at all, your remarks are very valuable and help to understand and develop the concept. Also saw over the past years some great posts, and you clearly know the physics very well.

I believe we agree that the best way forward would be to set up an experiment, as it may be hard or impossible to come to conclusions with only theoretical considerations. The "unknown knowns", like diffusion coefficients and gas transport are probably too complicated to figure out without experiment.

May I suggest a design of experiment as follows (for discussion), assuming we have "@PeerUnk magical device":

• we start with the simplest setup. A perfectly sealed tank (use duck tape) that has no leakage at all, and is filled with water, no plants. Water should not be stagnant, gentle surface agitation and turbulence in the top layer. Use @PeerUnk magical device to start injecting CO2 above the water, and stabilise it at 1% CO2. We could use a drop checker or pH probe to check the CO2 that diffuses and finally gets stabilised in the water. The pH curve of the water can easily be calculated back to understand the injection rate from "CO2 supplemented air" into the water.
• If we would find that the injection rate into the water, as you say @hax47 we have only 1% CO2 above, is too low, than we may change the setup so that the injection in not above the water, but with a diffuser in the water. @PeerUnk magical device will still set the CO2 target above the water at 1%, but in this case we have a more powerful injection mechanism using pure CO2 and the diffuser bubbles. Alternative we use CO2 Spray bar, as also @Marcel G is using in his experiment tanks.
• Next, we may create a 1-5% opening in the tank lid. We won't create an active gas flow, allow only for natural "diffusion" but like to understand what would be the limits of tank sealing to a) have a stable system and b) don't waste too much CO2.
• Next, when we have it working on a clean water tank, we bring in plants and see how their CO2 uptake changes the picture.

Does this make sense?

 

[Yugang]

What I was trying to refer to in Fick's law is that the diffusion rate is directly proportional to the partial pressure difference. And there we have more than >=200x difference when we compare 1% CO2 to 100% CO2.

If in the equilibrium the partial pressure in both the water and "CO2 supplemented air" is P-CO2-eq then my intuition says that the gas diffusion from water with P-CO2-eq to a gas pocket with "zero" CO2 equals the inverse: The gas diffusion from a gas pocket with P-CO2-eq back into water with "zero" CO2.
The rationale is that diffusion constants are same in both situations, but that concentration gradient is exactly the inverse leading to an opposite direction of the diffusion yet at same rate of diffusion. Would you agree?

If my intuition and argumentation are correct, then that mitigates your concern. We'll see from the experiment 🙂

 

[hax47]

If in the equilibrium the partial pressure in both the water and "CO2 supplemented air" is P-CO2-eq then my intuition says that the gas diffusion from water with P-CO2-eq to a gas pocket with "zero" CO2 equals the inverse: The gas diffusion from a gas pocket with P-CO2-eq back into water with "zero" CO2.
The rationale is that diffusion constants are same in both situations, but that concentration gradient is exactly the inverse leading to an opposite direction of the diffusion yet at same rate of diffusion. Would you agree?

If my intuition and argumentation are correct, then that mitigates your concern. We'll see from the experiment 🙂

I agree, my intuition also says that the direction does not matter, but I don't see how that mitigates my concern 🙂

Anyway, I am curious to see the results if @PeerUnk is kind enough to do the experiments.

 

[Yugang]

I agree, my intuition also says that the direction does not matter, but I don't see how that mitigates my concern

This and next year I unfortunately can't spend much time on my hobby, so I can't work out all physics. If my intuition serves me well, the reasoning goes as follows:

It is pretty well established, and I have confirmed this with some experiments as well, that in a CO2 injected tank the plant CO2 uptake is much less than the outgassing of CO2 from water to atmosphere. If we follow the above logic, and assume that the outgassing in traditional setup (diffusion from water to air) equals the "ingassing" (diffusion from CO2 supplemented air into water) in our new method, then we may conclude that the "ingassing" is much larger than the plant uptake. This, if correct and verified by experiment, would mitigate your concern that the rate of CO2 diffusion from a 1% CO2 concentration may not be enough.

What remains is the concern of leakage from the tank's overhead CO2 supplemented air pocket into the atmosphere, which could cause an unexpected high CO2 consumption. I believe only experiments could give a good indication, as this is nearly impossible to calculate or model.

 

[Bradders]

I wish I had paid more attention to physics at school!

A couple of questions, based only on my learning understanding of this thread.

1. If the air-space is held at 0.5 %–1 % CO2, is that going to equal 20-30PPM in the water? (i.e what most planted tank people want).
2. At night, fish and bacteria still need oxygen. If the lid is nearly sealed (not totally sealed, I hope!!!), how will this method make sure dissolved oxygen doesn’t fall below ~6 mg/l?
3. Aren't there many variables that can affect this, like tank volume, surface agitation, light level, plant biomass etc., i.e. can we really assume uniformly that plant CO2 uptake is much less than the out-gassing of CO2 from water to the atmosphere for a heavily planted dutech aquaiurm versus an aquairum with all anubius?

 

[Yugang]

If the air-space is held at 0.5 %–1 % CO2, is that going to equal 20-30PPM in the water? (i.e what most planted tank people want).

The hobby is often confused about ppm's, so I prefer to avoid mentioning ppm's here. We generally advise a 1 pH drop from outgassed water. When water is "outgassed", that is understood to mean in equilibrium with 400 ppm atmospheric CO2. Injecting CO2 to result in a 1 pH drop, means a 10 fold increase of the CO2 concentration, which implies in equilibrium with 10*400 ppm = 4000 ppm = 0.4% CO2. For a 1.3 pH drop, that is 2 times more, we would beed 2*0.4=0.8 % CO2 above the tank.

At night, fish and bacteria still need oxygen. If the lid is nearly sealed (not totally sealed, I hope!!!), how will this method make sure dissolved oxygen doesn’t fall below ~6 mg/l?

This would need to be tested, and is expected to be a function of plant mass as well as live stock. I have experimented with a nearly perfectly sealed tank, and no problem at all. If this can be applied generally, and how much "leakage" the tank must have, we will know after more testing.

can we really assume uniformly that plant CO2 uptake is much less than the out-gassing of CO2 from water to the atmosphere for a heavily planted dutech aquaiurm versus an aquairum with all anubius

In my tank yes, the loss of CO2 to the atmosphere is much higher than plant uptake. Also here, we need more data from more tanks.

This project is in a feasibility study phase, we are trying to figure out if it works by experiment and probably some more tweaking of the setup. Sometimes you research for one idea, and out comes a different and better idea. I am quite confident the idea will work in a professional/laboratory setting, but my confidence level would currently be only 60-70% in a hobbyist (affordable, simple, yet robust) setting.

I hope this helps.

 

[Bradders]

This project is in a feasibility study phase, we are trying to figure out if it works by experiment and probably some more tweaking of the setup. Sometimes you research for one idea, and out comes a different and better idea. I am quite confident the idea will work in a professional/laboratory setting, but my confidence level would currently be only 60-70% in a hobbyist (affordable, simple, yet robust) setting.

Very interesting one indeed. I will follow with keen interest!

I would say (as a side note, and notwithstanding the above) that when we consider the question "what problem are we trying to solve for?" in terms of CO2 usage in the hobby, I think that whether you use a in-tank diffuser, a reactor or a inline diffuser, there is still some uncertainty for the hobbyist. After you obtain some experience (from research, trial-error) you then understand the KH/pH relationship to CO2, how to rig the units up, how to sense when something is off, how to understand that drop checkers are 2 hours delayed, why bubble sizes can matter, why pH probes are unreliable, why flow is important etc etc. In short, the sum of the parts that make the whole is what induces confusion and elevates risk. In my humble view, the future king of the CO2 market is the one that moves us closer to 'plug-and-play' where CO2 safety, efficiency and level management becomes 'simple'. Don't know how to do it of course, but if a child of 10 cant intuitively make it work then I would say that we are not there!

 

[Yugang]

the future king of the CO2 market is the one that moves us closer to 'plug-and-play'

This is exactly the dream I am chasing. Have a manufacturer integrate the technology in the tank, so that the user just takes a remote control, or iPhone app, and sets the desired CO2 for the tank and forget about all the complications that we still have today (water chemistry, pH probe calibration, dialling in CO2, etc).

If this concept works, we'll have our plug and play for CO2.

 

[Yugang]

Sometimes you research for one idea, and out comes a different and better idea.

I am adding one idea here. If we don't like the idea of a nearly closed aquarium, or it appears not to work as expected, we can just place an inverted beaker in the tank, with an air bubble and the detector. This functions as a professional CO2 meter, and is independent of water chemistry. We could inject in the tank, using a reactor, and measure the CO2 concentration in the air pocket in the beaker, and use that in a feedback loop to the injection. When we get the time constants in the system right, as the beaker will need some time to be in equilibrium with the water, we can use this system to control the CO2 in the tank, plug and play.

It is not my favorite option, but a logical extension of the previous concept, or a fall back scenario.

 

[Little]

I have a double lidded tank so wanted to note a few things.
Positives
1. CO2 injection felt like it needed noticeably less gas.
2. Evaporation was reduced to imperceptible levels.
3. A few places that dropped condensation damaged floating plants but they still thrived.
4. The tank remained free of external pollution e.g no cat hair etc…
5. Despite running an air driven sump (also under a lid) co2 levels remain stable. I assume a balance was reached with how air exits before co2 due to density outweighing it diffusing with the co2?
Negatives
1. In summer I found temperatures to rise excessively. The solution I put in place was PWM fans. Another solution would be needed if you want to retain a co2 layer. I only use the fans when temperature rises over 25.
2. The lids need to be opened for daily feeding.
3. The light intensity needs to be significantly higher due to the lid becoming layered in condensation and watermarks etc…

I’m excited to see how this project develops. I’m also in the mindset of it being a lead to more informed thinking once testing starts

 

[Yugang]

I have a double lidded tank so wanted to note a few things.
Positives
1. CO2 injection felt like it needed noticeably less gas.
2. Evaporation was reduced to imperceptible levels.
3. A few places that dropped condensation damaged floating plants but they still thrived.
4. The tank remained free of external pollution e.g no cat hair etc…
5. Despite running an air driven sump (also under a lid) co2 levels remain stable. I assume a balance was reached with how air exits before co2 due to density outweighing it diffusing with the co2?
Negatives
1. In summer I found temperatures to rise excessively. The solution I put in place was PWM fans. Another solution would be needed if you want to retain a co2 layer. I only use the fans when temperature rises over 25.
2. The lids need to be opened for daily feeding.
3. The light intensity needs to be significantly higher due to the lid becoming layered in condensation and watermarks etc…

I’m excited to see how this project develops. I’m also in the mindset of it being a lead to more informed thinking once testing starts

Similar experience here.

I started experimenting with nearly perfectly closed tank as I wanted to "hibernate" the system during a 2 months absence from home. Nearly no evaporation, indicating almost no gas exchange. Plants and livestock no problem at all. I have had my tank like this for more than 6 months, and indeed saved a lot on CO2.

The main problem I see for the detector is water condensation, so we would have to operate it at slightly elevated temperatures. When opening the tank for feeding or maintenance the CO2 enriched air escapes, but after closing the tank again we just have to supplement to 0.5 or 1%, which is not a lot.

 

[hax47]

If the air-space is held at 0.5 %–1 % CO2, is that going to equal 20-30PPM in the water? (i.e what most planted tank people want).

I've entered your 0.5% as CO2 in air, which will diffuse in water to 7.5 ppm or mg/L @ 25 degrees Celsius. Seems like a reasonable target.

This I about the right concentration in ppm.

The hobby is often confused about ppm's, so I prefer to avoid mentioning ppm's here. We generally advise a 1 pH drop from outgassed water. When water is "outgassed", that is understood to mean in equilibrium with 400 ppm atmospheric CO2. Injecting CO2 to result in a 1 pH drop, means a 10 fold increase of the CO2 concentration, which implies in equilibrium with 10*400 ppm = 4000 ppm = 0.4% CO2. For a 1.3 pH drop, that is 2 times more, we would beed 2*0.4=0.8 % CO2 above the tank.

I think the 1 pH drop comes from the assumption that the outgassed water contains 3 ppm CO2, so the 1 pH drop results in 30 ppm. In reality, water, completely equilibrated with outdoor air, contains about 0.6 ppm of CO2. My guess as to why 3 ppm is considered the outgassed CO2 concentration is that it is usually determined indoors with higher CO2 levels, and also because at the corresponding low partial pressures, the diffusion of CO2 becomes very slow and takes a long time until the water truly equilibrates; the one-day outgassing rule might not be enough. But I am also not sure if anyone needs 30 ppm CO2, especially in soft/acidic water; 15 ppm from the 1% CO2 should suffice.

At night, fish and bacteria still need oxygen. If the lid is nearly sealed (not totally sealed, I hope!!!), how will this method make sure dissolved oxygen doesn’t fall below ~6 mg/l?

I did some calculations here to see what happens in a completely sealed aquarium with CO2 and O2. One of the conclusions is that the headspace air volume has a huge impact on the O2 drop. Since the solubility of the O2 is very low, and there could be a lot of O2 in the headspace, it is basically acting as a O2 reservoir. I think the safe oxygen supply can be managed with the right headspace volume.

In my tank yes, the loss of CO2 to the atmosphere is much higher than plant uptake. Also here, we need more data from more tanks.

I am in the process of installing an industrial-grade pH monitor on my current aquarium. I could also put a CO2 sensor in the headspace, and with my setup (a single dose of CO2 in the morning, and declining levels during the day), it will be very easy to determine the exact CO2 loss and plant uptake at different partial pressure levels (in mg/l/hour units, for example). I could even seal the lid (more or less) with duct tape and measure the CO2 uptake rate when CO2 is injected above the water. I do not have a CO2 control device, but a single dose and the consequent declining concentration (and parallel pH change in the water) should give us an idea how what the CO2 uptake looks like at different partial pressures and how much is lost into the environment.

 

[hax47]

1. In summer I found temperatures to rise excessively. The solution I put in place was PWM fans.

I also use a lid and had the same problem and came to the same solution. But I only turn on the fans during the night to preserve the CO2 during the day. One side effect is that since the water becomes colder than the room, there is no condensation on the lid anymore. The evaporation is a problem, though.

 

[Yugang]

@hax47 I am reading the above, I believe you've got it all, experimentally as well as theoretically. I will try and contribute as much as I can, but my time will be limited as I have a couple of external demands this year.

I hope we can continue as a team with everyone who is interested and willing to contribute. If we could have a start with an experimental setup and @PeerUnk magical device, we may have an opportunity to bring a new innovation to the hobby. Perhaps we can make "plug and play CO2" a reality. That would be a nice planted tank hobby new year wish for 2026.

 

[PeerUnk]

Anyway, I am curious to see the results if @PeerUnk is kind enough to do the experiments.

Well, I am. But time is very limited the next 4 months. This week I have some spare time. First step was fixing the device as the oxygen sensor needed to be replaced, and had to develop some prototype casing.
After some dumb wiring error I made, the prototype is completed. Exception is:

• the water temperature sensor, need to address an issue with it.
• the fan seems to be dead somehow.

Apologies for the messy wiring, but I need some sign of life first:


And it lives again!


Next step will be a manual calibration and test for valid measurements.

 

[xlucine]

1. In summer I found temperatures to rise excessively. The solution I put in place was PWM fans. Another solution would be needed if you want to retain a co2 layer. I only use the fans when temperature rises over 25.

Fans to exchange the CO2-enriched air for fresh could also be useful if O2 levels were an issue in the tank overnight

The main problem I see for the detector is water condensation, so we would have to operate it at slightly elevated temperatures. When opening the tank for feeding or maintenance the CO2 enriched air escapes, but after closing the tank again we just have to supplement to 0.5 or 1%, which is not a lot.

Assuming 1.2 kg/m3 for air, then a 1.2x0.4x0.05 m air space would only have ~30g of air. Getting that to 1% CO2 would need ~0.3g of CO2; about 1p worth

 

[Yugang]

Assuming 1.2 kg/m3 for air, then a 1.2x0.4x0.05 m air space would only have ~30g of air. Getting that to 1% CO2 would need ~0.3g of CO2; about 1p worth

This also gives a feeling as to how concerned we should be with perfectly sealing the tank. If one full replenishment of the CO2 supplemented headspace of the tank costs 1 pence, we can probably afford a less than perfect sealing without spending too much on CO2 losses by leakage. Not exact science, but it is an interesting data point to have an intuition.