Silicon Tubing Oxygenation During Racking

[Holden Caulfield]

Hey All:

I have been using a 6 foot 3/8" ID silicon tube to close transfer beer from a stainless steel Brewbucket to a CO2 purged keg - for sanitation I boil it prior to using which is why I like it. I know that silicone tubing is notorious for oxygen permeability, but given the short time the racking process takes (about 20 mins), I never gave it a thought. The other day though, I started to think that given the movement of the beer through the tubing and the surface to volume ratio of the tubing to the beer, this could actually be oxidizing my beer during the kegging process. While me beer is very good, especially when comparing to any American Craft Brewer's, when I compare it to German brewers like Tucher, Bitburger, and Ayinger, I am quite humbled.

Any thoughts or better yet, real data, on what I may be doing to my beer during transfer with silicone tubing?

My problem may also be on the hot side as my setup and energy level doesn't really allow for good hot side low oxygen brewing but suffice it to say it pretty standard for a cooler mashtun immersion chiller guy - so not so bad.

Thanks in advance.

 

[tracer bullet]

It's been discussed, and it's rare to find factual information on it. Most guesses are that it's not a long enough time to be worried about.

There are materials i.e. EVA barrier that could do it, but they are small diameter so it depends if you're talking 1 gallon, 3 gallons, 5, 10, etc. to be moved. You could also consider some stainless piping and just have silicone at a few spots for the connections (cheaper) or all stainless of course (more money and less flexibility. Brewhardware lets you customize tubing, I've used it a few times (I set up a stainless blow-off tube for example).

 

[doug293cz]

What is the OD of your silicon tubing? Wall thickness makes a difference in the rate at which O2 can diffuse thru the tubing.

Just as a gut feeling, I don't think using silicon tubing for a short transfer should cause much of a problem. I'll think about trying to quantify how much O2 might actually get thru in 20 minutes.

Brew on

 

[AlexKay]

Thick wall tubing will slow oxygen diffusion, but there also will be a greater reservoir of oxygen already in the plastic. I bet it’s a wash.

Why not shuffle things around so you need less tubing to make the transfer? Short work with a scissors could cut the oxygen uptake by a factor of 2 or even 3.

 

[SanPancho]

actually engineer mod already discussed this with us earlier and insisted it simply wasnt possible. odd that now its up for debate. maybe the opinion changes depending on the "messenger" or something.

this is obvious, but the most important factor is time. if you're just racking off then its probably not a big deal. yes, i'd say 20mins of oxidation is very bad. right? but the actual liquid being subject to oxidation is ONLY the liquid in the tubing. for the sake of argument, let's say its half a pint. and that half pint probably only stays in the tubing for less than 5 seconds. so i cant imagine its really an issue with any noticeable affect.

the place where it can be significant is during whirlpool, and for some folks who do recirc during mash. in that case you're talking about a continuous loop so the time of recirc is effectively the time of oxidation, i.e. the whole 20mins. but your beer is on a one-way trip. and it winds up in a purged keg, so all good there.

as noted above you can just switch to an EVA or other barrier type tubing for probably less than 10bucks. problem solved.

well, now that i think about it, that's probably not really the culprit if you suspect o2 damage. yes, silicone is permeable. but i'd wager the bigger oxygen risk is your connections between tubing and barbs. silicone and worm drive clamps are well known to be places to get air leaks. its a goldilocks of adjustment, too loose and you get air, too tight and you mash the silicone and can get air bubbles.
any air leaks at those connnections would do way more damage by letting in oxygen vs what could possibly permeate the tubing. so yeah, another point on the board for getting some sort of barrier tubing in place. in that case the EVA tubing with the push connectors are usually a solid, reliable setup.

 

[brewbama]

Have you considered EVA tubing?

“EVABarrier double wall tubing is an excellent choice for all draft dispensing applications. The inner barrier is an excellent gas barrier, keeping oxygen out and CO2 in. The inner wall is also extremely smooth to prevent bacteria and other microbes from attaching to the inside surface. The outer barrier gives the tubing a decent stiffness while still remaining pliable.”

 

[VikeMan]

but the actual liquid being subject to oxidation is ONLY the liquid in the tubing. for the sake of argument, let's say its half a pint. and that half pint probably only stays in the tubing for less than 5 seconds. so i cant imagine its really an issue with any noticeable affect.

The amount of liquid in the tubing isn't really the important thing. What matters is how much O2 diffuses all the way through the barrier during the entire racking time. Hopefully, @doug293cz will do the math and figure out how much O2 is getting through.

My own gut feeling is that it's not a lot. OTOH, there are lots of little ways that wort and beer get oxidized, and they add up.

 

[doug293cz]

The amount of liquid in the tubing isn't really the important thing. What matters is how much O2 diffuses all the way through the barrier during the entire racking time.

Yes. And also the total beer volume that that O2 gets distributed throughout.

Brew on

 

[hotbeer]

Store all your silicone tubing in containers flooded with CO2. Eventually most the O2 will be used up from them.

But this comes under the "too much effort for the gain" by my mental assessment at this time.




Interesting just how permeable O2 is in silicone compared to other gases.

https://imageserv5.team-logic.com/mediaLibrary/99/D116_20Haibing_20Zhang_20et_20al.pdf
See table 2 at the top of pdf page 5.

 

[Holden Caulfield]

Thanks all for your replies.

What is the OD of your silicon tubing?

Not sure, but I would guess it's about 1/8" thickness around 1/4" ID so 1/2"

Why not shuffle things around so you need less tubing to make the transfer? Short work with a scissors could cut the oxygen uptake by a factor of 2 or even 3.

At minimum, I will do this. I would expect the drop in O2 to be proportional to percentage length reduced.

EVA barrier that could do it, but they are small diameter so it depends if you're talking 1 gallon, 3 gallons, 5, 10, etc. to be moved.

EVABarrier double wall tubing is an excellent choice for all draft dispensing applications. The inner barrier is an excellent gas barrier, keeping oxygen out and CO2 in. The inner wall is also extremely smooth to prevent bacteria and other microbes from attaching to the inside surface.

I use EVA barrier throughout my keezer due to its superior oxygen impermeability - it is so much better than everything else. I only used the silicone for racking due to the fact I could boil it prior to use. Given EVA's slickness and that I only make 5 gal batches, I think I will take this approach as rinsing and running a little homemade PBW should clean it satisfactorily. Let me know if you think otherwise? Just before usage, Starsan will be pump through it to sanitize.

The amount of liquid in the tubing isn't really the important thing. What matters is how much O2 diffuses all the way through the barrier during the entire racking time.

At first this was my thinking, but then it occurred to me that the entire volume is pushed through it. Which means that every bit of the beer experiences the surface to volume ratio of tubing to liquid for the time spent in the tube. As the beer in the tube probably never achieves partial pressure equilibrium the actual amount of oxygen could be significantly less than the fully oxygenated volume of tube or significantly more.


Thanks again for all your replies. At this time, I will start by cutting my tube to minimum length required, and then move to EVA. Just need to figure out how to attach it the Brewbucket ball lock valve barb.

 

[RM-MN]

Are you brewing NEIPA's or IPA's where the tiniest bit of oxygen can change the shelf life of the beer? If not, you are worrying over very little.

 

[tracer bullet]

Are you brewing NEIPA's or IPA's where the tiniest bit of oxygen can change the shelf life of the beer? If not, you are worrying over very little.

It's a good point. I'm not convinced a little O2 is a bad thing actually, for some beers. Stouts sitting around for a year, etc. I would easily believe some of their changing and perceived improvement is due to it. I wouldn't do it intentionally but have long suspected it's a part of it.

 

[Teufelhunde]

Damn....and I thought I had a tendency to over think things.........

 

[doug293cz]

Edit: The following calculations contain a significant logical error which makes the result too low by a factor of about 5. The discussions following this post describe the problem. Corrected spreadsheets are here. doug293cz


Finally got around to doing some calculations of total O2 diffusion thru silicone tubing during a transfer. The calculation used Equation 4 from here, and used the value for O2 silicone rubber permeability from the same source. I ignored the O2 that is most likely already dissolved in the tubing walls, and only included the diffusion thru the tubing, assuming a static gradient of O2 thru the tubing. O2 partial pressure outside the tubing is assumed 21% of atmospheric pressure, and inside the tubing partial pressure is assumed as 0. If I didn't make any mistakes, the ppb of additional O2 from the transfer (assuming the input values shown below) came out at about 70 ppb. This is enough to be of concern for hoppy beers, especially NEIPAs (assuming the beer starts off with some low level of dissolved O2.)

I have attached copies of the calculation speadsheet (in two formats) for you to play with, and check the math if you are a nerd.

Brew on

 

[VikeMan]

If I didn't make any mistakes, the ppb of additional O2 from the transfer (assuming the input values shown below) came out at about 70 ppb. This is enough to be of concern for hoppy beers, especially NEIPAs.

Thanks for the calcs!

More than I would have SWAGed, but not totally surprising. There's a reason silicone is used for blood oxygenation membranes.

 

[Holden Caulfield]

If I didn't make any mistakes, the ppb of additional O2 from the transfer (assuming the input values shown below) came out at about 70 ppb.

^You are awesome!!!!!

Thanks - my concern was warranted. Despite the volume of tube being small, I failed to consider, until the other day, that the entire 5 gallons is transferred through the tubing so huge surface to volume ratio. So glad I asked the question.

Thankfully, there is an easy fix, just use EVA, which I have on hand, and I will use the adapter below to splice a 2" piece of the silicon tubing to the EVA tubing. This should result in only about a 1/4" of silicon hose that contacts beer so PPM should be reduced to almost 0.

Thanks again!!!!

 

[IslandLizard]

Thankfully, there is an easy fix, just use EVA, which I have on hand,

Will you be using 3/8" (9.5 mm) ID EVA tubing?

 

[tracer bullet]

I think you're on the right track subbing silicone hose if the O2 is a concern. You could also speed it up. I'm curious about the 20 minutes thing. Try the wort on a countertop and the keg on the floor if it's gravity fed.

Other thoughts -

1) If you are pushing CO2 into the fermenter and purging out the keg, you could probably speed that up. Dont' want it to blow apart but could speed it up.

2) If you are doing a closed loop gravity thing (it's what I do), keep in mind the return hose for the CO2 can allow O2 to come in. I wont' claim how much it matters or not, but if you're doing one side consider the other as well.

 

[PCABrewing]

Brew on

My only question is the time for transfer.
I haven't measured it yet but I think my gravity transfers (5 gal.) are complete in ~5 min.
I suppose that could be optimistic, I need to time one next time if I remember.
But otherwise thanks for doing the think-work.

 

[doug293cz]

My only question is the time for transfer.
I haven't measured it yet but I think my gravity transfers (5 gal.) are complete in ~5 min.
I suppose that could be optimistic, I need to time one next time if I remember.
But otherwise thanks for doing the think-work.

I just used numbers similar to what OP stated for their case. Since I included the spreadsheets that allow you to put in whatever numbers are appropriate for your case, you can run your own calc. Just be sure to use measured values, not estimates.

Brew on

 

[doug293cz]

Finally got around to doing some calculations of total O2 diffusion thru silicone tubing during a transfer. The calculation used Equation 4 from here, and used the value for O2 silicone rubber permeability from the same source. I ignored the O2 that is most likely already dissolved in the tubing walls, and only included the diffusion thru the tubing, assuming a static gradient of O2 thru the tubing. O2 partial pressure outside the tubing is assumed 21% of atmospheric pressure, and inside the tubing partial pressure is assumed as 0. If I didn't make any mistakes, the ppb of additional O2 from the transfer (assuming the input values shown below) came out at about 70 ppb. This is enough to be of concern for hoppy beers, especially NEIPAs (assuming the beer starts off with some low level of dissolved O2.)

I have attached copies of the calculation speadsheet (in two formats) for you to play with, and check the math if you are a nerd.

View attachment 808615
View attachment 808616


Brew on

There is one major uncertainty in how I did the calculations. The equation for amount of gas that diffuses thru the tubing wall gives the answer in "volume." But, is that volume at STP (0°C & 1 atm) or at the temp and partial pressure (P1) of the diffusing gas? I assumed the latter, and that's why I corrected back to 1 atm volume (4.776 cm^3 becomes 1.003 cm^3), before converting from volume to mols, and then mols to mass. But if the equation gives volume at STP, then that adjustment doesn't take place (and the molar volume also changes a little), and the the ppb ends up at about 360 ppm, and that would be way too much. Anyone have any thoughts on what the correct interpretation is?

Brew on

 

[AlexKay]

There is one major uncertainty in how I did the calculations. The equation for amount of gas that diffuses thru the tubing wall gives the answer in "volume." But, is that volume at STP (0°C & 1 atm) or at the temp and partial pressure (P1) of the diffusing gas? I assumed the latter, and that's why I corrected back to 1 atm volume (4.776 cm^3 becomes 1.003 cm^3), before converting from volume to mols, and then mols to mass. But if the equation gives volume at STP, then that adjustment doesn't take place (and the molar volume also changes a little), and the the ppb ends up at about 360 ppm, and that would be way too much. Anyone have any thoughts on what the correct interpretation is?

Brew on

You’ve already got partial pressure in your numerator. Assuming there isn’t a cooperative effect (seems reasonable!) you wouldn’t want to include it again.

 

[doug293cz]

You’ve already got partial pressure in your numerator. Assuming there isn’t a cooperative effect (seems reasonable!) you wouldn’t want to include it again.

The first occurrence relates to the driving force for diffusion, which is proportional to the difference in the partial pressures of the diffusing species on either side of the diffusion barrier.

The second is for adjustment of volume at one pressure to volume at another pressure,

The two are actually unrelated. I'll admit that I myself had to ponder whether or not a "double count" was going on.

It's the second occurrence that I am uncertain about whether it is required or not. I'm going to see if I can find permeability data that directly uses mass or mols of the diffusing species, rather than volume (which is too damn fungible.) Better units would remove the ambiguity.

Brew on

 

[AlexKay]

The first occurrence relates to the driving force for diffusion, which is proportional to the difference in the partial pressures of the diffusing species on either side of the diffusion barrier.

The second is for adjustment of volume at one pressure to volume at another pressure,

The two are actually unrelated. I'll admit that I myself had to ponder whether or not a "double count" was going on.

It's the second occurrence that I am uncertain about whether it is required or not. I'm going to see if I can find permeability data that directly uses mass or mols of the diffusing species, rather than volume (which is too damn fungible.) Better units would remove the ambiguity.

Brew on

At the atomistic scale, there is no driving force for diffusion. The movement across a gradient is statistical.

Maybe I’m oversimplifying, but my thinking is that in a given amount of time, a molecule has a X% chance of making it through. That should give a linear dependence on the partial pressure difference.

 

[doug293cz]

At the atomistic scale, there is no driving force for diffusion. The movement across a gradient is statistical.

Maybe I’m oversimplifying, but my thinking is that in a given amount of time, a molecule has a X% chance of making it through. That should give a linear dependence on the partial pressure difference.

You are correct at a rigorous level. I probably should have put "driving force" in quotations (like here .) At a phenomenalistic level, the statistics of movement in the presence of a gradient behave like a "driving force." The statistics relate to the entropy term in the GIbbs free energy, so we could frame it as force times distance = energy change (I'm pretty sure there is no net enthalpy change due to just the diffusion [although it's been a while. I'm trying to flex some brain cells I haven't used in a long time.])

Brew on

 

[AlexKay]

You are correct at a rigorous level. I probably should have put "driving force" in quotations (like here .) At a phenomenalistic level, the statistics of movement in the presence of a gradient behave like a "driving force." The statistics relate to the entropy term in the GIbbs free energy, so we could frame it as force times distance = energy change (I'm pretty sure there is no net enthalpy change due to just the diffusion [although it's been a while. I'm trying to flex some brain cells I haven't used in a long time.])

Brew on

My larger point is that if there are no interactions between molecules, you pretty much have to have a linear dependence on the number of molecules involved. It seems like if you count volume as "volume at a partial pressure," you end up getting a quadratic dependence.

 

[doug293cz]

My larger point is that if there are no interactions between molecules, you pretty much have to have a linear dependence on the number of molecules involved. It seems like if you count volume as "volume at a partial pressure," you end up getting a quadratic dependence.

I'll have to think about that.

Edit: "Volume at partial pressure" is a measure of the number of molecules involved: PV=nRT. So, that would argue that in the equation, volume is "volume at partial pressure" in order for the relationship to be linear. Thus to convert "volume at partial pressure" to mass, you need to convert volume at partial to volume at some other pressure for which you know the molar volume, which is what I did in the original calculation. Or, you could convert molar volume at 1 atm to molar volume at whatever the partial pressure is - the math is equivalent.

Brew on

 

[AlexKay]

I'll have to think about that.

Edit: "Volume at partial pressure" is a measure of the number of molecules involved: PV=nRT. So, that would argue that in the equation, volume is "volume at partial pressure" in order for the relationship to be linear. Thus to convert "volume at partial pressure" to mass, you need to convert volume at partial to volume at some other pressure for which you know the molar volume, which is what I did in the original calculation. Or, you could convert molar volume at 1 atm to molar volume at whatever the partial pressure is - the math is equivalent.

Brew on

But you've already got partial pressure in your "volume of O2" equation: permeability * area * time * partial pressure/thickness. But you're the one with the spreadsheet! If you double the partial pressure, does the ppb O2 go up by 2 or 4? (I agree, btw, that having a permeability with units that give you a volume is totally not helpful; witness our confusion.)

 

[maryc27182]

If you have time to do more research, it would be worth considering the transient aspects involved here. We probably should consider the lag time before steady-state is achieved, because if our transfer time is below the lag time number, I shouldn't think we transfer significant mass during the transient (gut feeling only--I have not done the calculation, nor has the single source I glanced at been run experimentally at home-brew-like conditions.)

Equation 4 you cite is for planar sheet, steady-state diffusion. The appropriate equations for a hollow cylindrical tube are easily found (both steady-state and transient), though at this level of approximation, the planar sheet is probably OK (but maybe not the assumption of steady-state flow.)

I think it would be worth checking to see if the lag times are greater than the typical home-brew transfer time before doing any more detailed analysis. My gut says that the lag time associated with transient diffusion through home-brew scale cylindrical tubing would be of the same order of magnitude--and therefore we don't make it to steady-state during transfer--but my gut is sometimes wrong.

Here's someone's dissertation from 1964 (which is surprisingly searchable!) and it has some experimental data which is at lower temps than is useful for us, but which gives some idea of the magnitude of lag times. Diffusion and Solution in Silicone Elastomers, H.T. Chio, 1964, Imperial College, London.

 

[AlexKay]

If you have time to do more research, it would be worth considering the transient aspects involved here. We probably should consider the lag time before steady-state is achieved, because if our transfer time is below the lag time number, I shouldn't think we transfer significant mass during the transient (gut feeling only--I have not done the calculation, nor has the single source I glanced at been run experimentally at home-brew-like conditions.)

The silicone is likely saturated with oxygen before you even start. Oxygen from outside will take some time to diffuse in, but the oxygen dissolved in the rubber will be going into solution at this time.