Sciencemadness Discussion Board

Control of reflux temperature for fractional distillation - how?

Sulaiman - 12-7-2023 at 06:46

I have not measured this, but I suspect that,
the liquid returned to the top of a fractionating column by the condenser
will be significantly cooler than its b.p.
This must make the top of the column less efficient than it could be.

I have seen diagrams of industrial fractionating columns where
the condensed vapour is returned to the column via a temperature controlled tank.
That makes product take-off easy and maintains fractionating power.
Is there a simple, practical method of doing something equivalent with my NS24 glassware?

PS I'm thinking more of some method of heating the liquid by the up-going vapour
rather than electrically etc.

The simplest alternative is to add more plates,
or just accept non-ideal performance.

[Edited on 13-7-2023 by Sulaiman]

Lionel Spanner - 12-7-2023 at 07:35

Industrial fractionating columns are designed for continuous processes; lab-scale equipment is better suited to batch processes, and attempting to construct continuous-process equipment at that scale is a very complicated task.

In batch processes, the distillate can be manually returned to the heating flask as often as necessary to refine it, and at a lab scale, this is not a practical problem. Doing this with multiple tons of liquid is a different story, which is why continuous processes, with multiple, temperature-controlled take-off points, are preferred in industry.

Also, at lab scale, a condenser is more typically attached to the outlet of the column rather than the top, so the mixture is not under reflux.

[Edited on 12-7-2023 by Lionel Spanner]

Rainwater - 12-7-2023 at 17:43

Quote: Originally posted by Sulaiman  
I have not measured this, but I suspect that,
the liquid returned to the top of a fractionating column by the condenser
will be significantly cooler than its b.p.
This must make the top of the column less efficient than it could be.
[Edited on 12-7-2023 by Sulaiman]

Assuming you can set up for total reflux and have some control of the takeoff
Then, that statement is true.

While playing with my appratus I noticed that chilled condencer coolant was breaking my equilibrium and causing occillation in my distillate tempature, The first method tried was is to adjust the flow rate of coolent until about 1/2 of the reflux column was wet. Cheap easy and it worked.

I wanted to do better, adding a thermometer into the coolant return line and adjusted the flow until the coolant returning is about 5-15c below the target head temperature, resulted in 1/3 to 1/2 of the reflux condencer being wet.

It also saves a lot of ice.

Sulaiman - 12-7-2023 at 20:57

So far I have been controlling take-off rate (hence reflux ratio)
by varying the cooling water flowrate to the reflux condenser,
allowing some of the vapour to pass through the reflux condenser to the product condenser.
Like this
and this

I can't recommend this method - too sensitive to coolant temperature, and flowrate
requiring frequent tweaking,
but on the positive side, it is easy to implement with common glassware,
and, relevant to this discussion,
the reflux condenser is by design/intent not over-cooling the condensed vapour,
so that's good.
(possibly a reason for tails compression as reported on the homedistiller page pointed to above? )

So, any other suggestions regarding equalisation of the temperature of the liquid returned to the top of the column by the reflux condenser?

I may eventually buy a partial-take-off head,
but this effect will still be of interest to me.

teodor - 13-7-2023 at 02:10

I did several fractionations with with partial-take-off head and if I put the reflux ration less than 1:5 there is no stable equilibrium in the column.
Probably you can put something like this part
20230713_120752.jpg - 3.7MB
between the top of the column and the condenser (which you put above, not into the side arm as usuall) and try to regulate the reflux ration just by regulating the heat (having a good insulation of the system up to the branch point). With the special head you have just a better/faster adjustment, but I think the partial take-off can work with this simple branch also. Also, I think if you have the condenser above the branch point it should be enough for not-too volatile liquids and you don't need the second condenser in the side arm.
The key idea is to keep the main condensation point just ~1cm above the branch, not inside the condenser. The liquid which flows down is heated by vapours and vapours are condensed here because the liquid takes their heat, so the top of the column (where the insulation starts, just below the branch) will have the temperature almost equal to the boiling point.

You still need the condenser on top, it will create the temperature gradient but will not actually condense the vapours except the most volatile part of them.

[Edited on 13-7-2023 by teodor]

Sulaiman - 13-7-2023 at 02:40

I've considered similar
(Vapor Management in home distiller terminology)
but with the addition of some kind of valve
to control reflux ratio.
(eg a disc shaped stirbar/stirdisc with an external magnet)

Endo - 13-7-2023 at 05:48

I have controlled takeoff by using a reflux condenser as a fractionation column, I feed water into the reflux column with a peristaltic pump from a sink full of cool water. As the temperature at the head increases I just increase the flow through the reflux column a bit to keep it steady and pull the head temp back down. The reflux ratio changes with each adjustment and by the end of the distillation the takeoff usually decreases.

I know it is not a hands free process, and depending on what is being distilled it can require many adjustments, but for me it works well. I often struggle to get good separations even with azeotropes because I am at 7300ft elevation.

Distillation.jpg - 2.2MB

Sulaiman - 13-7-2023 at 06:57

Quote: Originally posted by Endo  
I have controlled takeoff by using a reflux condenser as a fractionation column.....

Any idea approximately how many theoretical plates this achieves?

Endo - 18-7-2023 at 07:57

Sorry Sulaiman,

If I am consistent babysitting it and maintaining the head temp solidly at a good temp by adjusting the peristaltic pump at a good rate I typically just capture the fraction I need until the takeoff dwindles to nothing and at that point the higher boiling compounds end up at reflux.

FrecherChemiker - 19-7-2023 at 06:29

tbh I have never seen a setup with column + condenser on top of it.
Usually you control the destillation by heating and/or vacuum.
If separation is still not sufficient either use a larger column or distill multiple times.

Rainwater - 19-7-2023 at 16:09

So today i was forced to disolve and amount of 7,9-Dihydro-1H-purine-2,6,8(3H)-trione in about 8l of water, so while i waited for that, I hobbled over to my lab and setup a condencer controlled reflux like pictured above.

It can be built with simple cheap parts most labs have available

It is not as easy as a variable takeoff head.

The main downside is the difficulty in controlling the reflux, i had better results flowing the coolant backwards through the condencer.

Coolent flow rate was very tricky, but I must say. Once i swapped over from an inline valve to a more controlled wedge valve. I was impressed.
Then a thermometer into the coolant return line really helped.

If you can afford the loss in column length, and the extra time babysitting. This may be a cheaper option than purchasing a variable takeoff head

yobbo II - 21-7-2023 at 11:46

If you are using a hotplate could you use an external temperature sensor attached to condenser top ? You need an external temp. Input on hotplate.

There may be 'hunting' problems. The solution sounds too simple

[Edited on 21-7-2023 by yobbo II]

Rainwater - 21-7-2023 at 15:23

Ive tried that and having the controlling probe in the distillation head has its pros and cons.

On startup you risk full power being applied to your flask and uncontrolled boiling while the controller tries to figure things out.
The mantel will be adjusting the power for you, so as the composition changes, the boiling points change and the controller is lost at what to do.

In short, you have to move it around alot to get good seperation and fast distillation rates

Without tinkering it will go between maintaining the heat temp without producing distillate, and the distillate tempature oscillating providing poor seperation.

To start the distillation
1) Place the probe in the boiling flask and slowly increasing the temperature until the solution starts to boil.

2) As the solution gets depleted in volatile components, the distillate will come over less and less until it stops.

3)the mantel power will need to be regularly increased. For that reason, set the temperature controller to the current fraction tempature and move the controlling probe to the distillate head.

4) the distillate will start coming over but know that it will begin slowing down over time because of the way the controller works. As long as your not transitioning between fractions, momentarily, turning up the temp, or removing the probe will provide the feedback the controller needs to restore the distillation rate.

5) when you approach the end of a fraction, the constantly changing head temp will throw the controller into oscillations. Move the probe back into the boiling flask, then adjust the controller temp to the flask current tempature.

6) to finish pushing through a transition keep adjusting the controller until the head temp becomes stable then goto step 3

As far as how this relates to the above constructed appratus, you'll be adjusting 2 sensitive and touchy controls at the same time upping the level of difficulty. Not saying its. Bad Idea, just staying its going to be a little harder

BromicAcid - 21-7-2023 at 20:40

If you're doing a good fractional distillation you need a variable takeoff head to allow return to the column in order to get the proper reflux ratio and get efficient separation. Even if you have the packing, height and diameter to get 8 theoretical plates if you're just flashing material through the column with no return you're not going to get all those theoretical plates working as they should. Having been in the industry for awhile this is the head of choice. Obviously you need the right column for your separation but this gives rock-solid head temperatures reflecting the current boiling point of whatever component or mixture you're dealing with. This specific head is usually used for vacuum distillation as you can release the receiver to pressure and swap it out while leaving the pot under vacuum and not having to disrupt the equilibrium in the column.

Anyway, there's a needle valve that adjusts the takeoff rate, you can watch the return rate based on the drips off the condenser. Yes the material is chilled when dripping off the condenser but the thermometer is in the headspace so it reflects the actual boiling point. This is the setup I would emulate providing I had a good column.

Sulaiman - 22-7-2023 at 00:29

That distillation head looks like a rather expensive
'washing up' accident waiting to happen ;)

BromicAcid - 22-7-2023 at 04:46

That was just my lead-in to this one being half the price


Just that without a variable takeoff rate and being able to put the column in steady-state you're not going to be able to achieve maximum separation with your distillation column.

Rainwater - 22-7-2023 at 10:15

Quote: Originally posted by BromicAcid

If you're doing a good fractional distillation you need a variable takeoff head

Agreed but for 2800usd. Ouch. Like that doesnt even fathom in my brain how that could cost that much.

[Edited on 22-7-2023 by Rainwater]

Screenshot_20230722_141405_Samsung Internet.jpg - 152kB

BromicAcid - 22-7-2023 at 18:08

There's gotta be some Bomex version or something out there though.

Fery - 22-7-2023 at 21:30

Quote: Originally posted by Sulaiman  
the liquid returned to the top of a fractionating column by the condenser
will be significantly cooler than its b.p.

The heat of evaporation / condensation is much more higher than specific heat capacity. E.g. for water specific heat capacity 4,186 J/g°C and heat of evaporation 2256 kJ for 1 kg at 100°C at atm. pressure, so heat of evaporation is more than 500 times higher than specific heat capacity.
The colder reflux liquid is quickly heated to the boiling point at the cost of small amount of vapor condensation. There is always some distance (at least few cm) between the cooling part of the condenser and the top of the filling in the column which is enough for thermal equalization when distilling laboratory amounts.

Dr.Bob - 2-8-2023 at 09:34

If people really want some fancy distillation heads, I have a variety. A few are pristine, a few have damage or missing stopcocks that need to be replaced, and others are just odd. But way cheaper than the ones shown, I am sure. I have all sorts of stills, heads, various DeanStark type things, and more. I am happy to sell them.

arkoma - 2-8-2023 at 13:10

With EtOH, from raw sugar wash (estimated 16-18% w/hydrometer), I can get 90% purity easily in one run. I had to learn to apply heat slowly and I use a 4 litre Erlenmeyer with a 300mm 24/29 air condenser with five "balls" of copper mesh spaced out in it, still head/thermometer up top, 105° adapter on that going into the top of a 300mm Graham condenser. I ease up the temp and when the vapor front climbs I pay attention and back the temp down a little. I like to run "low and slow" and keep my fractionating column ALMOST choked down with liquid returning to the flask. It's pretty efficient at giving me a good fraction, but you better pull up a chair because it ain't fast.

Sulaiman, I remember when your technology for distillation consisted of a long glass tube at an angle with a wet towel draped on it for cooling. Come a ways!

[Edited on 8-2-2023 by arkoma]

[Edited on 8-2-2023 by arkoma]

Grizli7 - 17-12-2023 at 05:59

Handbuch der Laboratoriumsdestillation (Krell, 1976) may be this help

Grizli7 - 17-12-2023 at 06:06

Sulaiman - 17-12-2023 at 06:39

Quote: Originally posted by arkoma  
With EtOH, from raw sugar wash (estimated 16-18% w/hydrometer), I can get 90% purity easily in one run.
you might benefit from a quick 'stripping run' (simple distillation) to get 'low wines' (typ. 40+%ABV)
before doing a fractional distillation.
The volume to go in the pot is significantly reduced (to about 1/3rd)
(or three times as much product per tedious fractional distillation)
and foaming during fractionation is almost eliminated.

Grizli7: thanks for the pdf. Concise and useful.

Fery: thanks, I didn't consider the huge ratio,
so the slightly over-cooled condensate from the reflux condenser,
falling on the top of the column,
is not as significant an imperfection as I was imagining..
- still not perfect ;)

[Edited on 17-12-2023 by Sulaiman]

Grizli7 - 18-12-2023 at 09:01

Quote: Originally posted by Sulaiman  
Quote: Originally posted by Endo  
I have controlled takeoff by using a reflux condenser as a fractionation column.....

Any idea approximately how many theoretical plates this achieves?


Fery - 18-12-2023 at 10:12

Everyone who needs good separation should use efficient and insulated column, steady heating source, adjustable reflux ratio distillation head.
I did an experiment with fermented apple cider and tried how much pure ethanol I can get. I synthesized a lot of ethyl esters with amazing scents and I hate commercial ethanol denatured with stinky butanone.
I did 2 simple distillations of fermented apple cider and got cca 80% ethanol where could be cca upto 1% methanol (from pectin, apples are reach source of pectin), ethyl acetate and higher chain alcohols. Then I performed last step which was distillation on column. Imagine that 2 liters in the distillation flask should evaporate and condense few times until all the lower boiling sideproducts pass from distilling flask into the colum and then to the top of the colum (methanol and then azeotrope ethyl acetate + ethanol + water). Thus you need very high reflux ratio, like 10 drops returning back into the distillation flask and 1 drop taken into the receiver and starting collecting distillate must be AFTER enough of time so the lower b.p. sideproducts have enough time to pass from distillation flask into the colum (the whole content in the distillation flask should evaporate and condense very likely more than 1 time because b.p. differences are only at scale of 10-20 C). If you start to receive the main fraction too early there are still lower b.p. sideproducts in the distillation flask! It just requires some time and patience and efficient apparatus.
I tested 1 m long Hempel column packed with Raschig rings, variable reflux ratio distillation head with reflux ratio 10:1. Final ethanol was very pure, methanol below 0,1%, ethyl acetate below 0,1% - my friend Bedlasky performed GC in his job in a factory where they produce a lot of products, one of them diethyloxalate so he controls ethanol they import into the factory and he wrote me that they never had such pure ethanol as I produced in home lab apparatus.
Here the experiment, there are plenty of photos with thermometer during time what happens and how long does every step last:
Put it simply, the main fraction was started to collect after 1 hour, heating mantle at the power just below flooding column. Every picture contains exact time in its name. It could be possible to approximate how much condensate returned back into the distillation flask during that 1 hour.