Reflux Splitting

[myles]

Re: Mr P's Bubble Bumper
Re: Meeting time Boka and Brain Guys

I just had a thought that might be of interest for those that wish to increase production rates and decrease run times with packed columns.
It is probably not new as it combines aspects of a whole load of threads including Mike's paper on vapour reflux, but I can’t remember seeing it.

I just mentioned in Mr P's thread that bubbles ignore the slots and use the puke gap when there is low ABV on the plate due to surface tension. This also correlates with why the packing dries out at the end of a run and the reflux channels instead of wicking into the packing.

Now I believe that 1 factor affecting run times with packed columns, is that effectively the length of the column reduces as the alcohol is depleted out of the system. I say this because as the wetting potential of the reflux diminishes so does the effectiveness of the packing. In fact once the packing starts to dry out it does sodd all.

Now going back to Mike's idea of introducing high proof liquid to the base of the column. I think I am correct in saying that he was using the idea to super-charge the column by reducing the temperature gradient over its length, by changing the composition of the vapour in it.

However, it might also be useful to return a portion of the reflux from the top of the column into the packing (say 1/3rd up from the boiler) for no reason other than to increase the wetting potential of the reflux. This MIGHT contribute to preventing the packing drying out and this would have to contribute to reduced run times.

I have mentioned before that I think reflux splitting is a good idea (especially with low porosity packing like SPP) but this might be another application where it has benefits but for a different reason. Not by changing the vapour composition, but by changing the wetting properties of the reflux which would usually be alcohol depleted by the time they get to the base of the column in the later sections of the run.

What do you think?

[Frosteecat]

Sounds good but if you are using an automic reflux return won't it be depleted when you are needing it at the end of the run? Or are you talking about retaining some higher ABV fluid and "introducing" it as needed??

Ironically, that same discussion in PP's thread got me wondering if altering pressure in all or part of your column would aid in the process as well...maybe we combine it and call it "pressure whetting"!!

[Bushman]

Kind of goes along the same lines as the flyingdutchman is advocating! One part has me questioning your meaning:

Now I believe that 1 factor affecting run times with packed columns, is that effectively the length of the column reduces as the alcohol is depleted out of the system. I say this because as the wetting potential of the reflux diminishes so does the effectiveness of the packing. In fact once the packing starts to dry out it does sodd all.
When you state the length of the column reduces as the alcohol is depleted are you referring to the amount of tails that can actually work it's way up the column as the abv is reduced during the run?


What do you think?

[myles]

No Bushman it is something even more fundemental than that. Towards the end of the run the temperature at the base of the column and in the headspace rises and the ABV of the vapour being produced drops. Eventually a point is reached where the packing holds no liquid.

The reflux tends to run in chanels down the wall of the column. This can be seen in some of the videos of glass columns. Also it has been documented by plate users that the lower plate dries out and holds no liquid.
If either the plate holds no liquid or that section of packing holds no liquid, then the effective length of the column has just been shortened by the dry section. The upper sections of the column are still functioning but the lower section is not.

Now this is an endgame scenario that does not suddenly happen. It has to be assumed that there is a gradual decrease in the effective use of the column over time. I have no idea when it starts to occur.
However, if we can take steps to do something that actually counteracts this process and keeps the full length of the column in effective use for longer, then surely that would increase the efficiency of the column over the length of the entire run.

Now I am not entirely sure when this starts to occur as until now I have never had visibility inside my column. Perhaps someone in the know could clarify this situation. When running a plated column, at what point in the run does the lowest plate start to loose its ability to hold liquid? Can anyone with a packed column with a sight glass near the base provide some information regarding when the packing starts to dry out?

[myles]

Sounds good but if you are using an automic reflux return won't it be depleted when you are needing it at the end of the run? Or are you talking about retaining some higher ABV fluid and "introducing" it as needed??

Ironically, that same discussion in PP's thread got me wondering if altering pressure in all or part of your column would aid in the process as well...maybe we combine it and call it "pressure whetting"!!

I am not too concerned about tails but am thinking more about what is hapening towards the end of the hearts collection phase.
At this point the reflux condenser is still producing high ABV reflux, but how far down the column can it travel before it is depleted of alcohol and all that remains is very low ABV "column bottoms". In this scenario how effective is the lower section of packing in fractionating the low ABV vapour that is still being produced by the boiler?

That low ABV vapour needs a wetted surface to be fractionated by, and if that wet surface is not available, it will not be fractionated until it gets to a point higher up the column where it is available. Reintroducing some of the reflux liquid lower down the column will allow it to travel further before it comes alcohol depleted. This would increase the length of packing that is supporting liquid and is available to fractionate the vapour that is still being produced.

Well that's what I think at the moment. I suppose it all depends on how the reflux liquid behaves. Ideally you wish it to shed the last of its alcohol just before it drops back into the boiler - but not when it is half way down the column.

[Frosteecat]

Does it have to be ABV rich? Could you put a simple "mist port" where you need it and just wet with a fine mist of H20?

[Bushman]

Thanks I'm with you, I need to go back and reread the threads all this is giving me goose bumps thinking about where this site is going with new technology ideas!

[myles]

Does it have to be ABV rich? Could you put a simple "mist port" where you need it and just wet with a fine mist of H20?

Good question and it takes us straight back to the surface tension issue and the "wetability" of the packing. High ABV liquid wicks better than water, but any system that ensures the packing does not dry out would help. You still have the redistillation issue though.

Part 1 is that some of the heat in the vapour is re-evapourating alcohol from the reflux liquid. If that alcohol is not there then it can't be re-evapourated.
Part 2 is that higher boiling point vapours are condensing into the liquid layer. No liquid layer - no condensation.

An H2O mist would help part 2 but not part 1. I am sure it would still improve efficiency though, just by helping to segregate the high boiling point volatiles and by increasing the length of packing that is actually in use.

[LWTCS]

Great post Myles.
All the stuff you talked about is what this design addresses.
Though my thought process originally was that I would want the packing to stay wetted longer by recycling to a point of optimum purity... To the point where the material passing through the packed section was so pure that it would eventually flash and then render the packing dry for the opposite reason....All this would happen prior to collecting,,,,presumably

Even if I could only get the thing to behave this way for a breif period of time,,,i would think that the visual reference would be a good indicator for when to collect at very rapid take off speeds.......presumably

There is also an accommodation for a downcomer to be added or removed as I am not firm on how the head unit should be designed at this point.

[LWTCS]

This is sorta where I'm at with that design.
Still need to get the drain /external downcomer installed. And the reflux condenser is missing but that is easy.

[Prairiepiss]

I posted this on my thread but thought it would be good here to.

Now that Myles mentioned it. The reflux coming from the packing did change as the ABV dropped. During most of the run when there was a concentrated amount of alcohol in it. It would run down the side walls of the ball. Or wetting it. As the alcohol was depleted it started dripping more. So it wasn't wetting the outer walls of the ball any more. It would just drip from the column to the plate.

[myles]

Nice There is more than 1 way to skin a cat.

That particular idea seems just right for those packings that appear to work without a requirement for forced reflux. Like SPP. These seem to be able to generate their own reflux to keep themselves wet.

I still don't understand HOW that damned SPP works or in fact how much you need. Just waiting for someone to put 6" of the stuff on top of a column of scrubbers and say
"see it STILL works!!!"

Thanks Mr P, being able to see inside the ball looks like it has all sorts of advantages. Something I am bearing in mind.

[mash rookie]

Great post Myles.
All the stuff you talked about is what this design addresses.
Though my thought process originally was that I would want the packing to stay wetted longer by recycling to a point of optimum purity... To the point where the material passing through the packed section was so pure that it would eventually flash and then render the packing dry for the opposite reason....All this would happen prior to collecting,,,,presumably

Even if I could only get the thing to behave this way for a breif period of time,,,i would think that the visual reference would be a good indicator for when to collect at very rapid take off speeds.......presumably

There is also an accommodation for a downcomer to be added or removed as I am not firm on how the head unit should be designed at this point.

Larry, I did not notice before that you were putting packing in the column. Are you attempting to seperate fractions there while raising ABV with your thumpers?

Ironically, that same discussion in PP's thread got me wondering if altering pressure in all or part of your column would aid in the process as well...maybe we combine it and call it "pressure whetting"!!

Frosty, Pressure will have a negative affect on fraction separation by keeping partial pressures compacted.

[LWTCS]

I'll be using tee sections with sight glasses at the base, middle and top of the column in order to observe what is happening and how behavior changes at each phase of the run.
This here is a much more accurate mock up of the intended design.
Pretty sure the packed section will get insulated. I'll likely use the 2" column rather than the 3" as 4 liters per hour with a 2" was my goal.

[Frosteecat]

Man, that is some solid work. We're approaching scale model refinery distillation it appears...

[LWTCS]

Are you attempting to seperate fractions there while raising ABV with your thumpers?

Yes Scotty that pretty much nails it...Just want the thumpers to do the grunt work and let the packed section do what packed section does best. That along with my approach of not returning any thing (water or alcohol) to the boiler.

I figure the liquid in the thumpers also does a better job of managing unneeded heat also.

[Prairiepiss]

4 lt an hour. At 95%ABV? Man I need more power.

I think you just set the mark to be hit Larry. Last one there is a rotten egg.

[Prairiepiss]

Are you attempting to seperate fractions there while raising ABV with your thumpers?

Yes Scotty that pretty much nails it...Just want the thumpers to do the grunt work and let the packed section do what packed section does best. That along with my approach of not returning any thing (water or alcohol) to the boiler.

I figure the liquid in the thumpers also does a better job of managing unneeded heat also.

Exactly. Exploit all resources available. Not being confined by one design or theory. But using them all to get the most you can.

[LWTCS]

I figure the liquid in the thumpers also does a better job of managing unneeded heat also.

So I figure if we can keep the micro thumper below the packed section as enriched as possible it will allow the column to not have to cope with un needed heat and thus free up the entire packed section to do the same type of phase change that happens more commonly at the top of a packed section............All speculation of course

[mash rookie]

I stated a few months ago when you first showed your rig that even a basic packed column could benefit greatly from a boiler mounted thumper.

Is your plan is to run in total reflux until all alcohol is in the system then take off in fractional method? It sounds like a great marriage of two types of stills.

[Prairiepiss]

My thinking is this. If we can get the ABV as high as we can. Then give it the tools and space. The fractions will work themselves out naturally.

The more water we can remove. The easier the fractions will separate. And the lower temps Larry talks about will help it even further.

My two nickels.

[Frosteecat]

Does it have to be ABV rich? Could you put a simple "mist port" where you need it and just wet with a fine mist of H20?

Good question and it takes us straight back to the surface tension issue and the "wetability" of the packing. High ABV liquid wicks better than water, but any system that ensures the packing does not dry out would help. You still have the redistillation issue though.

Part 1 is that some of the heat in the vapour is re-evapourating alcohol from the reflux liquid. If that alcohol is not there then it can't be re-evapourated.
Part 2 is that higher boiling point vapours are condensing into the liquid layer. No liquid layer - no condensation.

An H2O mist would help part 2 but not part 1. I am sure it would still improve efficiency though, just by helping to segregate the high boiling point volatiles and by increasing the length of packing that is actually in use.

The circular nature of distillation makes my head spin . Just make smaller controllable circles is what my gut tells me but I'm probably wrong.
I don't want to mess with a bunch of pressures and elbows and drains and such. I'm lazy and my fab skills are Cro-magnon at best... I'm just going to put some kind of centrifuge at the top to "toss" liquid back down the damn column!!!

[LWTCS]

I'm not nearly as interested in making vodka as such MR. I just want to pull azeotrope at 4 liters per hour. Only to make my peepee feel better about his misfortune.
Seriously, its just to see if I can do it.
This build is why I talked to you about your observations with draining down multiple plate levels. Evidently there were concerns that to drain that far down would be disruptive to the gradient .

[mash rookie]

Larry, I know you guys are chasing the ABV pretty hard. In practice the return drain did not affect temperature negatively as some suggested it might. In light of my better understanding about partial pressures and contamination thoughts, I am surprised it works as well as it does. Actually it is probably working better than expected but limited because of the contamination issue.

My thinking is this. If we can get the ABV as high as we can. Then give it the tools and space. The fractions will work themselves out naturally.

The more water we can remove. The easier the fractions will separate. And the lower temps Larry talks about will help it even further.

My two nickels.

PP, They are not just going to work them selves out unless you provide an opportunity for them to do so. Meaning that as well as the thumper works at separating water and raising ABV, you will still need a plated or packed column above it with sufficient plates or packing space to allow fractions to separate. Yes, they will likely separate easier once water is removed but likely limited in speed to the diameter of the column above.


I think we are getting way off topic discussing your thumper towers. Myles, I am not quit getting your concern about packing drying as compound pressures change. Other than redistilling some of the liquid, I not following what advantage it would serve to wet the packing by recycling distillate. I think your observations may cause us to consider different types, sizes and or density of packing at different levels. I think when you see liquid running down the sides but not on the packing you are seeing condensed vapor that would be making it higher up the column except for the passive reflux of the colder pipe wall. The changing wash and vapor composition as a run progresses and temperature rises dictates that condensation will not happen at the lower levels. Is it your desire to cool this area by wetting the packing? Now that higher BP vapor is coming off will you get better separation by getting it to condense at the bottom of the column?

[Prairiepiss]

My thinking is this. If we can get the ABV as high as we can. Then give it the tools and space. The fractions will work themselves out naturally.

The more water we can remove. The easier the fractions will separate. And the lower temps Larry talks about will help it even further.

My two nickels.

I think I said the same thing you just did MR.

[LWTCS]

What contamination issues are you refering to MR?

[myles]

I'm not nearly as interested in making vodka as such MR. I just want to pull azeotrope at 4 liters per hour. Only to make my peepee feel better about his misfortune.
Seriously, its just to see if I can do it.
This build is why I talked to you about your observations with draining down multiple plate levels. Evidently there were concerns that to drain that far down would be disruptive to the gradient .

Thats pushing limmits Larry. For each 4 litres of azeotrope taken there is another 1.3 litres falling out the bottom of the column. Thats 5.3 litres of condensate per hour or 88.3 ml per minute. Given that your condensers will produce 78.6ml / kW removed / minute, you would need to be putting 1.8kW into the column (not the boiler).

Now by my calculations that has pushed your vapour speed up to about 23" or 24" per second. It will be interesting to see if this is achievable, we take the old guideline of 20" per second for mesh/scrubbers for granted. Perhaps it is time to push a bit.

[myles]

Myles, I am not quit getting your concern about packing drying as compound pressures change. Other than redistilling some of the liquid, I not following what advantage it would serve to wet the packing by recycling distillate. I think your observations may cause us to consider different types, sizes and or density of packing at different levels. I think when you see liquid running down the sides but not on the packing you are seeing condensed vapor that would be making it higher up the column except for the passive reflux of the colder pipe wall. The changing wash and vapor composition as a run progresses and temperature rises dictates that condensation will not happen at the lower levels. Is it your desire to cool this area by wetting the packing? Now that higher BP vapor is coming off will you get better separation by getting it to condense at the bottom of the column?

I might be looking at this from a slightly different point of view.

If we start with a standard 48" x 2" column at the start of the run we are using all 48" of packing. I absolutely agree that the changing wash and vapour composition combined with the temperature rise, means that the condensation does not occur in the lower portion of packing. As a consequence (at a given moment during the run) we might only utilise - for example - the top 24" of packing. The lower 24" does nothing at all.

Now if we can reintroduce reflux into this area to wet the packing, this should FORCE a change in the temperature gradient. This should in turn bring more of the packing back into use. I suppose this could equate to comparing the performance of a 48" x 2" column and a 60" x 2" column.
Those are just picked at random, I am not trying to quantify the effect of reflux splitting. By compressing the gradient between the boiler and the EFFECTIVE base of the packed column, are we not simmulating the use of a taller column?

[Prairiepiss]

Now your talking like the old world class reflux cm still. That had the through tube down lower. Odin seems to have gotten his to run good.

I was actually thinking kinda like this. But more like two columns in one type thing. One feeding the other. I think its possible. But it would take some fine controls to pull it off. Along with adjustments as the run went on.

But then again I have anterior motives for such a device. That I haven't gotten into yet. Save that for another thread at another time.

[LWTCS]

For each 4 litres of azeotrope taken there is another 1.3 litres falling out the bottom of the column. Thats 5.3 litres of condensate per hour or 88.3 ml per minute. Given that your condensers will produce 78.6ml / kW removed / minute, you would need to be putting 1.8kW into the column (not the boiler).

Now by my calculations that has pushed your vapour speed up to about 23" or 24" per second. It will be interesting to see if this is achievable, we take the old guideline of 20" per second for mesh/scrubbers for granted. Perhaps it is time to push a bit.

My experiment is a bit of a cheat really Myles. I can't speak to any of the statistics that you reference...
I am effectively trying to do all or nearly all of the distilling before I collect one single drop,,,and also utilize rising boiler temps to feed the system the heat that is needed.
I am hoping that the system will not "starve" for alcohol like an over sized column on an under sized boiler,,,,as I am shifting as much alcohol upward as possible and hope that the radiant (or the like) heat will be enough to drive the parasitic boilers..( think how the cm's stop rendering product in spite of the continued in put of heat).

Did I say that clearly? Don't think those stats apply in this particular instance.